Datasets:

ftopal
/

huggingface-models

Dataset card Data Studio Files Files and versions Community

Split (1)

train · 320k rows

pipeline_tag stringclasses 48 values	library_name stringclasses 198 values	text stringlengths 1 900k	metadata stringlengths 2 438k	id stringlengths 5 122	last_modified null	tags sequencelengths 1 1.84k	sha null	created_at stringlengths 25 25	arxiv sequencelengths 0 201	languages sequencelengths 0 1.83k	tags_str stringlengths 17 9.34k	text_str stringlengths 0 389k	text_lists sequencelengths 0 722	processed_texts sequencelengths 1 723
image-classification	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # vit-base-patch16-224-dmae-va-U5-42D This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 1.0383 - Accuracy: 0.55 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.003 - train_batch_size: 32 - eval_batch_size: 32 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 128 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 42 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\| \| No log \| 0.9 \| 7 \| 1.0970 \| 0.5167 \| \| 1.3527 \| 1.94 \| 15 \| 1.0383 \| 0.55 \| \| 1.3527 \| 2.97 \| 23 \| 1.2351 \| 0.4167 \| \| 1.3013 \| 4.0 \| 31 \| 1.3025 \| 0.3333 \| \| 1.3706 \| 4.9 \| 38 \| 1.3800 \| 0.2167 \| \| 1.3706 \| 5.94 \| 46 \| 1.4609 \| 0.1833 \| \| 1.4415 \| 6.97 \| 54 \| 1.3718 \| 0.4333 \| \| 1.3602 \| 8.0 \| 62 \| 1.3173 \| 0.3167 \| \| 1.3602 \| 8.9 \| 69 \| 1.2827 \| 0.4 \| \| 1.3079 \| 9.94 \| 77 \| 1.3167 \| 0.3167 \| \| 1.3247 \| 10.97 \| 85 \| 1.2579 \| 0.4 \| \| 1.3247 \| 12.0 \| 93 \| 1.3202 \| 0.2 \| \| 1.3102 \| 12.9 \| 100 \| 1.2354 \| 0.45 \| \| 1.2807 \| 13.94 \| 108 \| 1.3610 \| 0.25 \| \| 1.2807 \| 14.97 \| 116 \| 1.2803 \| 0.4 \| \| 1.2774 \| 16.0 \| 124 \| 1.3338 \| 0.2167 \| \| 1.2774 \| 16.9 \| 131 \| 1.2549 \| 0.35 \| \| 1.2596 \| 17.94 \| 139 \| 1.2693 \| 0.3667 \| \| 1.2413 \| 18.97 \| 147 \| 1.3005 \| 0.2167 \| \| 1.2413 \| 20.0 \| 155 \| 1.2299 \| 0.4333 \| \| 1.262 \| 20.9 \| 162 \| 1.3454 \| 0.2667 \| \| 1.2261 \| 21.94 \| 170 \| 1.2818 \| 0.3167 \| \| 1.2261 \| 22.97 \| 178 \| 1.2498 \| 0.4333 \| \| 1.2405 \| 24.0 \| 186 \| 1.3376 \| 0.3167 \| \| 1.2245 \| 24.9 \| 193 \| 1.2595 \| 0.3667 \| \| 1.2245 \| 25.94 \| 201 \| 1.3319 \| 0.4 \| \| 1.2034 \| 26.97 \| 209 \| 1.2528 \| 0.3833 \| \| 1.1818 \| 28.0 \| 217 \| 1.3656 \| 0.3667 \| \| 1.1818 \| 28.9 \| 224 \| 1.2501 \| 0.3833 \| \| 1.1479 \| 29.94 \| 232 \| 1.3241 \| 0.3 \| \| 1.1193 \| 30.97 \| 240 \| 1.3803 \| 0.3667 \| \| 1.1193 \| 32.0 \| 248 \| 1.2294 \| 0.4167 \| \| 1.1071 \| 32.9 \| 255 \| 1.4134 \| 0.5 \| \| 1.1071 \| 33.94 \| 263 \| 1.4123 \| 0.3667 \| \| 1.0429 \| 34.97 \| 271 \| 1.2184 \| 0.5 \| \| 1.0528 \| 36.0 \| 279 \| 1.3100 \| 0.45 \| \| 1.0528 \| 36.9 \| 286 \| 1.3249 \| 0.3833 \| \| 1.0055 \| 37.94 \| 294 \| 1.3051 \| 0.5 \| ### Framework versions - Transformers 4.38.2 - Pytorch 2.2.1+cu121 - Datasets 2.18.0 - Tokenizers 0.15.2	{"license": "apache-2.0", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "google/vit-base-patch16-224", "model-index": [{"name": "vit-base-patch16-224-dmae-va-U5-42D", "results": []}]}	Augusto777/vit-base-patch16-224-dmae-va-U5-42D	null	[ "transformers", "tensorboard", "safetensors", "vit", "image-classification", "generated_from_trainer", "base_model:google/vit-base-patch16-224", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T02:39:15+00:00	[]	[]	TAGS #transformers #tensorboard #safetensors #vit #image-classification #generated_from_trainer #base_model-google/vit-base-patch16-224 #license-apache-2.0 #autotrain_compatible #endpoints_compatible #region-us	vit-base-patch16-224-dmae-va-U5-42D =================================== This model is a fine-tuned version of google/vit-base-patch16-224 on an unknown dataset. It achieves the following results on the evaluation set: * Loss: 1.0383 * Accuracy: 0.55 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * learning\_rate: 0.003 * train\_batch\_size: 32 * eval\_batch\_size: 32 * seed: 42 * gradient\_accumulation\_steps: 4 * total\_train\_batch\_size: 128 * optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 * lr\_scheduler\_type: linear * lr\_scheduler\_warmup\_ratio: 0.1 * num\_epochs: 42 ### Training results ### Framework versions * Transformers 4.38.2 * Pytorch 2.2.1+cu121 * Datasets 2.18.0 * Tokenizers 0.15.2	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 0.003\n* train\\_batch\\_size: 32\n* eval\\_batch\\_size: 32\n* seed: 42\n* gradient\\_accumulation\\_steps: 4\n* total\\_train\\_batch\\_size: 128\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* lr\\_scheduler\\_warmup\\_ratio: 0.1\n* num\\_epochs: 42", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* Pytorch 2.2.1+cu121\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #tensorboard #safetensors #vit #image-classification #generated_from_trainer #base_model-google/vit-base-patch16-224 #license-apache-2.0 #autotrain_compatible #endpoints_compatible #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 0.003\n* train\\_batch\\_size: 32\n* eval\\_batch\\_size: 32\n* seed: 42\n* gradient\\_accumulation\\_steps: 4\n* total\\_train\\_batch\\_size: 128\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* lr\\_scheduler\\_warmup\\_ratio: 0.1\n* num\\_epochs: 42", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* Pytorch 2.2.1+cu121\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]
text-generation	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	OwOOwO/dumbo-krillin61	null	[ "transformers", "safetensors", "llama", "text-generation", "conversational", "arxiv:1910.09700", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T02:39:33+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #llama #text-generation #conversational #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #llama #text-generation #conversational #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
text-generation	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	ZachX/UltraMerge-v2-7B	null	[ "transformers", "safetensors", "mistral", "text-generation", "arxiv:1910.09700", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "8-bit", "region:us" ]	null	2024-04-19T02:39:57+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #mistral #text-generation #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #text-generation-inference #8-bit #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #mistral #text-generation #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #text-generation-inference #8-bit #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
text2text-generation	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # led-pubmed-20k-4096 This model is a fine-tuned version of [pszemraj/led-base-book-summary](https://huggingface.co/pszemraj/led-base-book-summary) on the None dataset. It achieves the following results on the evaluation set: - Loss: 1.1066 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 1 - eval_batch_size: 1 - seed: 42 - gradient_accumulation_steps: 16 - total_train_batch_size: 16 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 500 - num_epochs: 1 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| \|:-------------:\|:-----:\|:----:\|:---------------:\| \| 1.22 \| 0.75 \| 500 \| 1.1066 \| ### Framework versions - Transformers 4.38.2 - Pytorch 2.2.1+cu121 - Datasets 2.18.0 - Tokenizers 0.15.2	{"license": "bsd-3-clause", "tags": ["generated_from_trainer"], "base_model": "pszemraj/led-base-book-summary", "model-index": [{"name": "led-pubmed-20k-4096", "results": []}]}	Chung-Fan/led-pubmed-20k-4096	null	[ "transformers", "safetensors", "led", "text2text-generation", "generated_from_trainer", "base_model:pszemraj/led-base-book-summary", "license:bsd-3-clause", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T02:42:04+00:00	[]	[]	TAGS #transformers #safetensors #led #text2text-generation #generated_from_trainer #base_model-pszemraj/led-base-book-summary #license-bsd-3-clause #autotrain_compatible #endpoints_compatible #region-us	led-pubmed-20k-4096 =================== This model is a fine-tuned version of pszemraj/led-base-book-summary on the None dataset. It achieves the following results on the evaluation set: * Loss: 1.1066 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * learning\_rate: 5e-05 * train\_batch\_size: 1 * eval\_batch\_size: 1 * seed: 42 * gradient\_accumulation\_steps: 16 * total\_train\_batch\_size: 16 * optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 * lr\_scheduler\_type: linear * lr\_scheduler\_warmup\_steps: 500 * num\_epochs: 1 ### Training results ### Framework versions * Transformers 4.38.2 * Pytorch 2.2.1+cu121 * Datasets 2.18.0 * Tokenizers 0.15.2	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 5e-05\n* train\\_batch\\_size: 1\n* eval\\_batch\\_size: 1\n* seed: 42\n* gradient\\_accumulation\\_steps: 16\n* total\\_train\\_batch\\_size: 16\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* lr\\_scheduler\\_warmup\\_steps: 500\n* num\\_epochs: 1", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* Pytorch 2.2.1+cu121\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #safetensors #led #text2text-generation #generated_from_trainer #base_model-pszemraj/led-base-book-summary #license-bsd-3-clause #autotrain_compatible #endpoints_compatible #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 5e-05\n* train\\_batch\\_size: 1\n* eval\\_batch\\_size: 1\n* seed: 42\n* gradient\\_accumulation\\_steps: 16\n* total\\_train\\_batch\\_size: 16\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* lr\\_scheduler\\_warmup\\_steps: 500\n* num\\_epochs: 1", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* Pytorch 2.2.1+cu121\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]
null	peft	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # phi-2-ipo-renew1 This model is a fine-tuned version of [lole25/phi-2-sft-ultrachat-lora](https://huggingface.co/lole25/phi-2-sft-ultrachat-lora) on the HuggingFaceH4/ultrafeedback_binarized dataset. It achieves the following results on the evaluation set: - Loss: 2028.0933 - Rewards/chosen: -0.1243 - Rewards/rejected: -0.2158 - Rewards/accuracies: 0.6900 - Rewards/margins: 0.0915 - Logps/rejected: -255.1287 - Logps/chosen: -269.0499 - Logits/rejected: 0.5909 - Logits/chosen: 0.5352 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-06 - train_batch_size: 4 - eval_batch_size: 4 - seed: 42 - distributed_type: multi-GPU - gradient_accumulation_steps: 4 - total_train_batch_size: 16 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 2 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Rewards/chosen \| Rewards/rejected \| Rewards/accuracies \| Rewards/margins \| Logps/rejected \| Logps/chosen \| Logits/rejected \| Logits/chosen \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------------:\|:----------------:\|:------------------:\|:---------------:\|:--------------:\|:------------:\|:---------------:\|:-------------:\| \| 2496.843 \| 0.05 \| 100 \| 2502.2668 \| -0.0003 \| -0.0002 \| 0.5005 \| -0.0002 \| -233.5649 \| -256.6506 \| 0.8888 \| 0.8318 \| \| 2499.2807 \| 0.1 \| 200 \| 2494.8354 \| 0.0001 \| -0.0005 \| 0.5190 \| 0.0006 \| -233.5995 \| -256.6106 \| 0.8882 \| 0.8310 \| \| 2477.7609 \| 0.16 \| 300 \| 2481.5015 \| -0.0011 \| -0.0031 \| 0.5595 \| 0.0019 \| -233.8548 \| -256.7285 \| 0.8892 \| 0.8319 \| \| 2428.4195 \| 0.21 \| 400 \| 2419.1045 \| -0.0068 \| -0.0156 \| 0.6495 \| 0.0089 \| -235.1127 \| -257.2951 \| 0.8983 \| 0.8404 \| \| 2296.8842 \| 0.26 \| 500 \| 2349.4358 \| -0.0240 \| -0.0419 \| 0.6565 \| 0.0179 \| -237.7379 \| -259.0124 \| 0.8806 \| 0.8214 \| \| 2254.5846 \| 0.31 \| 600 \| 2273.4993 \| -0.0525 \| -0.0829 \| 0.6570 \| 0.0304 \| -241.8383 \| -261.8659 \| 0.8478 \| 0.7868 \| \| 2330.7787 \| 0.37 \| 700 \| 2224.3350 \| -0.0819 \| -0.1221 \| 0.6630 \| 0.0402 \| -245.7631 \| -264.8093 \| 0.8128 \| 0.7517 \| \| 2223.6863 \| 0.42 \| 800 \| 2196.0991 \| -0.1009 \| -0.1487 \| 0.6675 \| 0.0478 \| -248.4222 \| -266.7057 \| 0.7611 \| 0.6992 \| \| 2066.7418 \| 0.47 \| 900 \| 2166.0732 \| -0.1112 \| -0.1658 \| 0.6700 \| 0.0546 \| -250.1319 \| -267.7397 \| 0.7518 \| 0.6917 \| \| 2119.2691 \| 0.52 \| 1000 \| 2138.9312 \| -0.1215 \| -0.1821 \| 0.6715 \| 0.0606 \| -251.7610 \| -268.7693 \| 0.7213 \| 0.6619 \| \| 2191.7109 \| 0.58 \| 1100 \| 2121.8115 \| -0.1257 \| -0.1906 \| 0.6695 \| 0.0648 \| -252.6059 \| -269.1910 \| 0.7176 \| 0.6584 \| \| 2308.1883 \| 0.63 \| 1200 \| 2110.3069 \| -0.1409 \| -0.2123 \| 0.6665 \| 0.0715 \| -254.7812 \| -270.7044 \| 0.6920 \| 0.6330 \| \| 1996.7178 \| 0.68 \| 1300 \| 2095.3130 \| -0.1314 \| -0.2042 \| 0.6755 \| 0.0728 \| -253.9726 \| -269.7621 \| 0.6722 \| 0.6141 \| \| 2038.3844 \| 0.73 \| 1400 \| 2085.0852 \| -0.1383 \| -0.2140 \| 0.6800 \| 0.0756 \| -254.9441 \| -270.4488 \| 0.6513 \| 0.5933 \| \| 2094.2182 \| 0.79 \| 1500 \| 2076.3042 \| -0.1390 \| -0.2166 \| 0.6790 \| 0.0777 \| -255.2133 \| -270.5129 \| 0.6474 \| 0.5898 \| \| 2171.3457 \| 0.84 \| 1600 \| 2069.3757 \| -0.1374 \| -0.2166 \| 0.6810 \| 0.0792 \| -255.2130 \| -270.3595 \| 0.6392 \| 0.5818 \| \| 2189.3863 \| 0.89 \| 1700 \| 2062.1995 \| -0.1386 \| -0.2192 \| 0.6780 \| 0.0806 \| -255.4675 \| -270.4739 \| 0.6291 \| 0.5723 \| \| 2292.8938 \| 0.94 \| 1800 \| 2053.1299 \| -0.1196 \| -0.2005 \| 0.6830 \| 0.0809 \| -253.6025 \| -268.5789 \| 0.6275 \| 0.5703 \| \| 2085.5805 \| 0.99 \| 1900 \| 2052.3237 \| -0.1086 \| -0.1906 \| 0.6900 \| 0.0821 \| -252.6131 \| -267.4730 \| 0.6319 \| 0.5747 \| \| 1847.759 \| 1.05 \| 2000 \| 2050.4177 \| -0.1118 \| -0.1953 \| 0.6850 \| 0.0836 \| -253.0827 \| -267.7950 \| 0.6333 \| 0.5763 \| \| 2024.9559 \| 1.1 \| 2100 \| 2046.7593 \| -0.1219 \| -0.2083 \| 0.6900 \| 0.0864 \| -254.3799 \| -268.8073 \| 0.6157 \| 0.5590 \| \| 2038.6354 \| 1.15 \| 2200 \| 2043.5728 \| -0.1205 \| -0.2072 \| 0.6880 \| 0.0867 \| -254.2731 \| -268.6722 \| 0.6083 \| 0.5518 \| \| 2022.9617 \| 1.2 \| 2300 \| 2035.5857 \| -0.1173 \| -0.2041 \| 0.6895 \| 0.0868 \| -253.9597 \| -268.3491 \| 0.6101 \| 0.5535 \| \| 1871.641 \| 1.26 \| 2400 \| 2036.3373 \| -0.1190 \| -0.2073 \| 0.6895 \| 0.0884 \| -254.2831 \| -268.5161 \| 0.6046 \| 0.5482 \| \| 1907.3463 \| 1.31 \| 2500 \| 2034.7010 \| -0.1216 \| -0.2108 \| 0.6880 \| 0.0892 \| -254.6297 \| -268.7765 \| 0.6022 \| 0.5460 \| \| 1884.6086 \| 1.36 \| 2600 \| 2033.7977 \| -0.1215 \| -0.2105 \| 0.6910 \| 0.0890 \| -254.6014 \| -268.7708 \| 0.6013 \| 0.5451 \| \| 2034.9129 \| 1.41 \| 2700 \| 2032.5447 \| -0.1235 \| -0.2140 \| 0.6900 \| 0.0905 \| -254.9471 \| -268.9633 \| 0.5987 \| 0.5426 \| \| 2068.2822 \| 1.47 \| 2800 \| 2030.8698 \| -0.1251 \| -0.2162 \| 0.6900 \| 0.0911 \| -255.1671 \| -269.1270 \| 0.5943 \| 0.5383 \| \| 1977.4029 \| 1.52 \| 2900 \| 2030.6033 \| -0.1251 \| -0.2162 \| 0.6895 \| 0.0911 \| -255.1690 \| -269.1252 \| 0.5941 \| 0.5381 \| \| 2110.2887 \| 1.57 \| 3000 \| 2030.5707 \| -0.1259 \| -0.2173 \| 0.6905 \| 0.0915 \| -255.2821 \| -269.2050 \| 0.5908 \| 0.5348 \| \| 2068.2863 \| 1.62 \| 3100 \| 2029.4174 \| -0.1242 \| -0.2156 \| 0.6935 \| 0.0914 \| -255.1087 \| -269.0390 \| 0.5913 \| 0.5357 \| \| 1977.8852 \| 1.67 \| 3200 \| 2026.1289 \| -0.1249 \| -0.2165 \| 0.6960 \| 0.0916 \| -255.2016 \| -269.1071 \| 0.5920 \| 0.5364 \| \| 2123.3787 \| 1.73 \| 3300 \| 2027.3552 \| -0.1248 \| -0.2162 \| 0.6930 \| 0.0914 \| -255.1666 \| -269.0933 \| 0.5926 \| 0.5370 \| \| 1945.4934 \| 1.78 \| 3400 \| 2025.7804 \| -0.1248 \| -0.2164 \| 0.6935 \| 0.0916 \| -255.1899 \| -269.1010 \| 0.5909 \| 0.5353 \| \| 1937.2627 \| 1.83 \| 3500 \| 2027.8240 \| -0.1247 \| -0.2163 \| 0.6930 \| 0.0916 \| -255.1750 \| -269.0878 \| 0.5903 \| 0.5347 \| \| 2007.2062 \| 1.88 \| 3600 \| 2025.3228 \| -0.1244 \| -0.2164 \| 0.6895 \| 0.0919 \| -255.1843 \| -269.0623 \| 0.5910 \| 0.5352 \| \| 2076.715 \| 1.94 \| 3700 \| 2027.4857 \| -0.1243 \| -0.2159 \| 0.6920 \| 0.0916 \| -255.1383 \| -269.0487 \| 0.5913 \| 0.5358 \| \| 2055.2201 \| 1.99 \| 3800 \| 2027.8082 \| -0.1244 \| -0.2160 \| 0.6920 \| 0.0916 \| -255.1455 \| -269.0543 \| 0.5902 \| 0.5347 \| ### Framework versions - PEFT 0.7.1 - Transformers 4.36.2 - Pytorch 2.1.2 - Datasets 2.14.6 - Tokenizers 0.15.2	{"license": "mit", "library_name": "peft", "tags": ["alignment-handbook", "generated_from_trainer", "trl", "dpo"], "datasets": ["HuggingFaceH4/ultrafeedback_binarized"], "base_model": "microsoft/phi-2", "model-index": [{"name": "phi-2-ipo-renew1", "results": []}]}	DUAL-GPO-2/phi-2-ipo-renew1	null	[ "peft", "safetensors", "phi", "alignment-handbook", "generated_from_trainer", "trl", "dpo", "custom_code", "dataset:HuggingFaceH4/ultrafeedback_binarized", "base_model:microsoft/phi-2", "license:mit", "region:us" ]	null	2024-04-19T02:43:27+00:00	[]	[]	TAGS #peft #safetensors #phi #alignment-handbook #generated_from_trainer #trl #dpo #custom_code #dataset-HuggingFaceH4/ultrafeedback_binarized #base_model-microsoft/phi-2 #license-mit #region-us	phi-2-ipo-renew1 ================ This model is a fine-tuned version of lole25/phi-2-sft-ultrachat-lora on the HuggingFaceH4/ultrafeedback\_binarized dataset. It achieves the following results on the evaluation set: * Loss: 2028.0933 * Rewards/chosen: -0.1243 * Rewards/rejected: -0.2158 * Rewards/accuracies: 0.6900 * Rewards/margins: 0.0915 * Logps/rejected: -255.1287 * Logps/chosen: -269.0499 * Logits/rejected: 0.5909 * Logits/chosen: 0.5352 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * learning\_rate: 5e-06 * train\_batch\_size: 4 * eval\_batch\_size: 4 * seed: 42 * distributed\_type: multi-GPU * gradient\_accumulation\_steps: 4 * total\_train\_batch\_size: 16 * optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 * lr\_scheduler\_type: cosine * lr\_scheduler\_warmup\_ratio: 0.1 * num\_epochs: 2 ### Training results ### Framework versions * PEFT 0.7.1 * Transformers 4.36.2 * Pytorch 2.1.2 * Datasets 2.14.6 * Tokenizers 0.15.2	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 5e-06\n* train\\_batch\\_size: 4\n* eval\\_batch\\_size: 4\n* seed: 42\n* distributed\\_type: multi-GPU\n* gradient\\_accumulation\\_steps: 4\n* total\\_train\\_batch\\_size: 16\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: cosine\n* lr\\_scheduler\\_warmup\\_ratio: 0.1\n* num\\_epochs: 2", "### Training results", "### Framework versions\n\n\n* PEFT 0.7.1\n* Transformers 4.36.2\n* Pytorch 2.1.2\n* Datasets 2.14.6\n* Tokenizers 0.15.2" ]	[ "TAGS\n#peft #safetensors #phi #alignment-handbook #generated_from_trainer #trl #dpo #custom_code #dataset-HuggingFaceH4/ultrafeedback_binarized #base_model-microsoft/phi-2 #license-mit #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 5e-06\n* train\\_batch\\_size: 4\n* eval\\_batch\\_size: 4\n* seed: 42\n* distributed\\_type: multi-GPU\n* gradient\\_accumulation\\_steps: 4\n* total\\_train\\_batch\\_size: 16\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: cosine\n* lr\\_scheduler\\_warmup\\_ratio: 0.1\n* num\\_epochs: 2", "### Training results", "### Framework versions\n\n\n* PEFT 0.7.1\n* Transformers 4.36.2\n* Pytorch 2.1.2\n* Datasets 2.14.6\n* Tokenizers 0.15.2" ]
fill-mask	transformers	<!-- This model card has been generated automatically according to the information Keras had access to. You should probably proofread and complete it, then remove this comment. --> # jfforero/distilbert-base-uncased-BERT-POET This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset. It achieves the following results on the evaluation set: - Train Loss: 5.3153 - Validation Loss: 4.9360 - Epoch: 0 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'module': 'transformers.optimization_tf', 'class_name': 'WarmUp', 'config': {'initial_learning_rate': 2e-05, 'decay_schedule_fn': {'module': 'keras.optimizers.schedules', 'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': -998, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered_name': None}, 'warmup_steps': 1000, 'power': 1.0, 'name': None}, 'registered_name': 'WarmUp'}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.01} - training_precision: float32 ### Training results \| Train Loss \| Validation Loss \| Epoch \| \|:----------:\|:---------------:\|:-----:\| \| 5.3153 \| 4.9360 \| 0 \| ### Framework versions - Transformers 4.38.2 - TensorFlow 2.15.0 - Datasets 2.18.0 - Tokenizers 0.15.2	{"license": "apache-2.0", "tags": ["generated_from_keras_callback"], "base_model": "distilbert-base-uncased", "model-index": [{"name": "jfforero/distilbert-base-uncased-BERT-POET", "results": []}]}	jfforero/distilbert-base-uncased-BERT-POET	null	[ "transformers", "tf", "distilbert", "fill-mask", "generated_from_keras_callback", "base_model:distilbert-base-uncased", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T02:44:40+00:00	[]	[]	TAGS #transformers #tf #distilbert #fill-mask #generated_from_keras_callback #base_model-distilbert-base-uncased #license-apache-2.0 #autotrain_compatible #endpoints_compatible #region-us	jfforero/distilbert-base-uncased-BERT-POET ========================================== This model is a fine-tuned version of distilbert-base-uncased on an unknown dataset. It achieves the following results on the evaluation set: * Train Loss: 5.3153 * Validation Loss: 4.9360 * Epoch: 0 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * optimizer: {'name': 'AdamWeightDecay', 'learning\_rate': {'module': 'transformers.optimization\_tf', 'class\_name': 'WarmUp', 'config': {'initial\_learning\_rate': 2e-05, 'decay\_schedule\_fn': {'module': 'keras.optimizers.schedules', 'class\_name': 'PolynomialDecay', 'config': {'initial\_learning\_rate': 2e-05, 'decay\_steps': -998, 'end\_learning\_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered\_name': None}, 'warmup\_steps': 1000, 'power': 1.0, 'name': None}, 'registered\_name': 'WarmUp'}, 'decay': 0.0, 'beta\_1': 0.9, 'beta\_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight\_decay\_rate': 0.01} * training\_precision: float32 ### Training results ### Framework versions * Transformers 4.38.2 * TensorFlow 2.15.0 * Datasets 2.18.0 * Tokenizers 0.15.2	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* optimizer: {'name': 'AdamWeightDecay', 'learning\\_rate': {'module': 'transformers.optimization\\_tf', 'class\\_name': 'WarmUp', 'config': {'initial\\_learning\\_rate': 2e-05, 'decay\\_schedule\\_fn': {'module': 'keras.optimizers.schedules', 'class\\_name': 'PolynomialDecay', 'config': {'initial\\_learning\\_rate': 2e-05, 'decay\\_steps': -998, 'end\\_learning\\_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered\\_name': None}, 'warmup\\_steps': 1000, 'power': 1.0, 'name': None}, 'registered\\_name': 'WarmUp'}, 'decay': 0.0, 'beta\\_1': 0.9, 'beta\\_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight\\_decay\\_rate': 0.01}\n* training\\_precision: float32", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* TensorFlow 2.15.0\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #tf #distilbert #fill-mask #generated_from_keras_callback #base_model-distilbert-base-uncased #license-apache-2.0 #autotrain_compatible #endpoints_compatible #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* optimizer: {'name': 'AdamWeightDecay', 'learning\\_rate': {'module': 'transformers.optimization\\_tf', 'class\\_name': 'WarmUp', 'config': {'initial\\_learning\\_rate': 2e-05, 'decay\\_schedule\\_fn': {'module': 'keras.optimizers.schedules', 'class\\_name': 'PolynomialDecay', 'config': {'initial\\_learning\\_rate': 2e-05, 'decay\\_steps': -998, 'end\\_learning\\_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered\\_name': None}, 'warmup\\_steps': 1000, 'power': 1.0, 'name': None}, 'registered\\_name': 'WarmUp'}, 'decay': 0.0, 'beta\\_1': 0.9, 'beta\\_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight\\_decay\\_rate': 0.01}\n* training\\_precision: float32", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* TensorFlow 2.15.0\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]
null	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # augmented_mamba This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set: - Loss: 0.3991 - Accuracy: 0.9233 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 4 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - lr_scheduler_warmup_ratio: 0.01 - num_epochs: 2 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\| \| 0.0741 \| 0.2 \| 558 \| 0.3561 \| 0.8943 \| \| 0.0036 \| 0.4 \| 1116 \| 0.3148 \| 0.9090 \| \| 0.0001 \| 0.6 \| 1674 \| 0.4026 \| 0.92 \| \| 0.0006 \| 0.8 \| 2232 \| 0.2829 \| 0.9257 \| \| 0.003 \| 1.0 \| 2790 \| 0.3330 \| 0.9210 \| \| 0.0008 \| 1.2 \| 3348 \| 0.4015 \| 0.9238 \| \| 0.0002 \| 1.4 \| 3906 \| 0.4544 \| 0.9210 \| \| 0.0012 \| 1.6 \| 4464 \| 0.4062 \| 0.9262 \| \| 1.6264 \| 1.8 \| 5022 \| 0.3991 \| 0.9233 \| ### Framework versions - Transformers 4.38.2 - Pytorch 2.2.1+cu121 - Datasets 2.18.0 - Tokenizers 0.15.2	{"tags": ["generated_from_trainer"], "metrics": ["accuracy"], "model-index": [{"name": "augmented_mamba", "results": []}]}	erostrate9/augmented_mamba	null	[ "transformers", "pytorch", "generated_from_trainer", "endpoints_compatible", "region:us" ]	null	2024-04-19T02:48:22+00:00	[]	[]	TAGS #transformers #pytorch #generated_from_trainer #endpoints_compatible #region-us	augmented\_mamba ================ This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set: * Loss: 0.3991 * Accuracy: 0.9233 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * learning\_rate: 5e-05 * train\_batch\_size: 4 * eval\_batch\_size: 16 * seed: 42 * optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 * lr\_scheduler\_type: cosine * lr\_scheduler\_warmup\_ratio: 0.01 * num\_epochs: 2 ### Training results ### Framework versions * Transformers 4.38.2 * Pytorch 2.2.1+cu121 * Datasets 2.18.0 * Tokenizers 0.15.2	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 5e-05\n* train\\_batch\\_size: 4\n* eval\\_batch\\_size: 16\n* seed: 42\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: cosine\n* lr\\_scheduler\\_warmup\\_ratio: 0.01\n* num\\_epochs: 2", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* Pytorch 2.2.1+cu121\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #pytorch #generated_from_trainer #endpoints_compatible #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 5e-05\n* train\\_batch\\_size: 4\n* eval\\_batch\\_size: 16\n* seed: 42\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: cosine\n* lr\\_scheduler\\_warmup\\_ratio: 0.01\n* num\\_epochs: 2", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* Pytorch 2.2.1+cu121\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]
null	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	HC-85/distilbert-qlora-arxiv-multilabel	null	[ "transformers", "safetensors", "arxiv:1910.09700", "endpoints_compatible", "region:us" ]	null	2024-04-19T02:48:44+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #arxiv-1910.09700 #endpoints_compatible #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #arxiv-1910.09700 #endpoints_compatible #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
fill-mask	transformers	# 5CD-AI/viso-twhin-bert-large ## Overview <!-- Provide a quick summary of what the model is/does. --> We reduce TwHIN-BERT's vocabulary size to 20k on the UIT dataset and continue pretraining for 10 epochs. Here are the results on 4 downstream tasks on Vietnamese social media texts, including Emotion Recognition(UIT-VSMEC), Hate Speech Detection(UIT-HSD), Spam Reviews Detection(ViSpamReviews), Hate Speech Spans Detection(ViHOS): <table> <tr align="center"> <td rowspan=2><b>Model</td> <td rowspan=2><b>Avg</td> <td colspan=3><b>Emotion Recognition</td> <td colspan=3><b>Hate Speech Detection</td> <td colspan=3><b>Spam Reviews Detection</td> <td colspan=3><b>Hate Speech Spans Detection</td> </tr> <tr align="center"> <td><b>Acc</td> <td><b>WF1</td> <td><b>MF1</td> <td><b>Acc</td> <td><b>WF1</td> <td><b>MF1</td> <td><b>Acc</td> <td><b>WF1</td> <td><b>MF1</td> <td><b>Acc</td> <td><b>WF1</td> <td><b>MF1</td> </tr> <tr align="center"> <td align="left">viBERT</td> <td>78.16</td> <td>61.91</td> <td>61.98</td> <td>59.7</td> <td>85.34</td> <td>85.01</td> <td>62.07</td> <td>89.93</td> <td>89.79</td> <td>76.8</td> <td>90.42</td> <td>90.45</td> <td>84.55</td> </tr> <tr align="center"> <td align="left">vELECTRA</td> <td>79.23</td> <td>64.79</td> <td>64.71</td> <td>61.95</td> <td>86.96</td> <td>86.37</td> <td>63.95</td> <td>89.83</td> <td>89.68</td> <td>76.23</td> <td>90.59</td> <td>90.58</td> <td>85.12</td> </tr> <tr align="center"> <td align="left">PhoBERT-Base </td> <td>79.3</td> <td>63.49</td> <td>63.36</td> <td>61.41</td> <td>87.12</td> <td>86.81</td> <td>65.01</td> <td>89.83</td> <td>89.75</td> <td>76.18</td> <td>91.32</td> <td>91.38</td> <td>85.92</td> </tr> <tr align="center"> <td align="left">PhoBERT-Large</td> <td>79.82</td> <td>64.71</td> <td>64.66</td> <td>62.55</td> <td>87.32</td> <td>86.98</td> <td>65.14</td> <td>90.12</td> <td>90.03</td> <td>76.88</td> <td>91.44</td> <td>91.46</td> <td>86.56</td> </tr> <tr align="center"> <td align="left">ViSoBERT</td> <td>81.58</td> <td>68.1</td> <td>68.37</td> <td>65.88</td> <td>88.51</td> <td>88.31</td> <td>68.77</td> <td>90.99</td> <td>90.92</td> <td>79.06</td> <td>91.62</td> <td>91.57</td> <td>86.8</td> </tr> <tr align="center"> <td align="left">visobert-14gb-corpus</td> <td>82.2</td> <td>68.69</td> <td>68.75</td> <td>66.03</td> <td>88.79</td> <td>88.6</td> <td>69.57</td> <td>91.02</td> <td>90.88</td> <td>77.13</td> <td>93.69</td> <td>93.63</td> <td>89.66</td> </tr> <tr align="center"> <td align="left">viso-twhin-bert-large</td> <td><b>83.87</td> <td><b>73.45</td> <td><b>73.14</td> <td><b>70.99</td> <td><b>88.86</td> <td><b>88.8</td> <td><b>70.81</td> <td><b>91.6</td> <td><b>91.47</td> <td><b>79.07</td> <td><b>94.08</td> <td><b>93.96</td> <td><b>90.22</td> </tr> </div> </table> ## Usage (HuggingFace Transformers) Install `transformers` package: pip install transformers Then you can use this model for fill-mask task like this: ```python from transformers import pipeline model_path = "5CD-AI/viso-twhin-bert-large" mask_filler = pipeline("fill-mask", model_path) mask_filler("đúng nhận sai <mask>", top_k=10) ``` ## Fine-tune Configuration We fine-tune `5CD-AI/viso-twhin-bert-large` on 4 downstream tasks with `transformer` library with the following configuration: - train_batch_size: 16 - seed: 42 - gradient_accumulation_steps: 4 - weight_decay: 0.01 - optimizer: AdamW with betas=(0.9, 0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - training_epochs: 30 - model_max_length: 128 - metric_for_best_model: wf1 - strategy: epoch And different additional configurations for each task: \| Emotion Recognition \| Hate Speech Detection \| Spam Reviews Detection \| Hate Speech Spans Detection \| \| --------------------------------------------------------------------------------- \| --------------------------------------------------------------------------------- \| --------------------------------------------------------------------------------- \| --------------------------------------------------------------------------------- \| \|\- learning_rate: 1e-5\| \- learning_rate: 5e-6 \| \- learning_rate: 1e-5 \| \- learning_rate: 5e-6 \|	{"library_name": "transformers", "tags": [], "pipeline_tag": "fill-mask", "widget": [{"text": "shop l\u00e0m \u0103n nh\u01b0 c\u00e1i <mask>"}, {"text": "hag t\u1eeb Qu\u1ea3ng <mask> k\u1ef1c n\u00e9t"}, {"text": "Set xinh qu\u00e1, <mask> b\u00e8o nh\u00e8o"}, {"text": "\u0111\u00fang nh\u1eadn sai <mask>"}]}	5CD-AI/viso-twhin-bert-large	null	[ "transformers", "safetensors", "bert", "fill-mask", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T02:50:52+00:00	[]	[]	TAGS #transformers #safetensors #bert #fill-mask #autotrain_compatible #endpoints_compatible #region-us	5CD-AI/viso-twhin-bert-large ============================ Overview -------- We reduce TwHIN-BERT's vocabulary size to 20k on the UIT dataset and continue pretraining for 10 epochs. Here are the results on 4 downstream tasks on Vietnamese social media texts, including Emotion Recognition(UIT-VSMEC), Hate Speech Detection(UIT-HSD), Spam Reviews Detection(ViSpamReviews), Hate Speech Spans Detection(ViHOS): Usage (HuggingFace Transformers) -------------------------------- Install 'transformers' package: ``` pip install transformers ``` Then you can use this model for fill-mask task like this: Fine-tune Configuration ----------------------- We fine-tune '5CD-AI/viso-twhin-bert-large' on 4 downstream tasks with 'transformer' library with the following configuration: * train\_batch\_size: 16 * seed: 42 * gradient\_accumulation\_steps: 4 * weight\_decay: 0.01 * optimizer: AdamW with betas=(0.9, 0.999) and epsilon=1e-08 * lr\_scheduler\_type: cosine * training\_epochs: 30 * model\_max\_length: 128 * metric\_for\_best\_model: wf1 * strategy: epoch And different additional configurations for each task:	[]	[ "TAGS\n#transformers #safetensors #bert #fill-mask #autotrain_compatible #endpoints_compatible #region-us \n" ]
text-generation	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	12thD/I-SOLAR-10.7B-dpo-sft-v0.2	null	[ "transformers", "safetensors", "llama", "text-generation", "arxiv:1910.09700", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T02:50:58+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #llama #text-generation #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #llama #text-generation #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
text-generation	null	## Model Details Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety. Model developers Meta Variations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants. Input Models input text only. Output Models generate text and code only. Model Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. <table> <tr> <td> </td> <td><strong>Training Data</strong> </td> <td><strong>Params</strong> </td> <td><strong>Context length</strong> </td> <td><strong>GQA</strong> </td> <td><strong>Token count</strong> </td> <td><strong>Knowledge cutoff</strong> </td> </tr> <tr> <td rowspan="2" >Llama 3 </td> <td rowspan="2" >A new mix of publicly available online data. </td> <td>8B </td> <td>8k </td> <td>Yes </td> <td rowspan="2" >15T+ </td> <td>March, 2023 </td> </tr> <tr> <td>70B </td> <td>8k </td> <td>Yes </td> <td>December, 2023 </td> </tr> </table> Llama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability. Model Release Date April 18, 2024. Status This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback. License A custom commercial license is available at: [https://llama.meta.com/llama3/license](https://llama.meta.com/llama3/license) Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model [README](https://github.com/meta-llama/llama3). For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go [here](https://github.com/meta-llama/llama-recipes). ## Intended Use Intended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. Out-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English. Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. ## How to use This repository contains two versions of Meta-Llama-3-8B-Instruct, for use with transformers and with the original `llama3` codebase. ### Use with transformers See the snippet below for usage with Transformers: ```python >>> import transformers >>> import torch >>> model_id = "meta-llama/Meta-Llama-3-70B" >>> pipeline = transformers.pipeline( "text-generation", model=model_id, model_kwargs={"torch_dtype": torch.bfloat16}, device_map="auto" ) >>> pipeline("Hey how are you doing today?") ``` ### Use with `llama3` Please, follow the instructions in the [repository](https://github.com/meta-llama/llama3). To download Original checkpoints, see the example command below leveraging `huggingface-cli`: ``` huggingface-cli download meta-llama/Meta-Llama-3-70B --include "original/" --local-dir Meta-Llama-3-70B ``` For Hugging Face support, we recommend using transformers or TGI, but a similar command works. ## Hardware and Software Training Factors* We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute. Carbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program. <table> <tr> <td> </td> <td><strong>Time (GPU hours)</strong> </td> <td><strong>Power Consumption (W)</strong> </td> <td><strong>Carbon Emitted(tCO2eq)</strong> </td> </tr> <tr> <td>Llama 3 8B </td> <td>1.3M </td> <td>700 </td> <td>390 </td> </tr> <tr> <td>Llama 3 70B </td> <td>6.4M </td> <td>700 </td> <td>1900 </td> </tr> <tr> <td>Total </td> <td>7.7M </td> <td> </td> <td>2290 </td> </tr> </table> CO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others. ## Training Data Overview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data. Data Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively. ## Benchmarks In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see [here](https://github.com/meta-llama/llama3/blob/main/eval_methodology.md). ### Base pretrained models <table> <tr> <td><strong>Category</strong> </td> <td><strong>Benchmark</strong> </td> <td><strong>Llama 3 8B</strong> </td> <td><strong>Llama2 7B</strong> </td> <td><strong>Llama2 13B</strong> </td> <td><strong>Llama 3 70B</strong> </td> <td><strong>Llama2 70B</strong> </td> </tr> <tr> <td rowspan="6" >General </td> <td>MMLU (5-shot) </td> <td>66.6 </td> <td>45.7 </td> <td>53.8 </td> <td>79.5 </td> <td>69.7 </td> </tr> <tr> <td>AGIEval English (3-5 shot) </td> <td>45.9 </td> <td>28.8 </td> <td>38.7 </td> <td>63.0 </td> <td>54.8 </td> </tr> <tr> <td>CommonSenseQA (7-shot) </td> <td>72.6 </td> <td>57.6 </td> <td>67.6 </td> <td>83.8 </td> <td>78.7 </td> </tr> <tr> <td>Winogrande (5-shot) </td> <td>76.1 </td> <td>73.3 </td> <td>75.4 </td> <td>83.1 </td> <td>81.8 </td> </tr> <tr> <td>BIG-Bench Hard (3-shot, CoT) </td> <td>61.1 </td> <td>38.1 </td> <td>47.0 </td> <td>81.3 </td> <td>65.7 </td> </tr> <tr> <td>ARC-Challenge (25-shot) </td> <td>78.6 </td> <td>53.7 </td> <td>67.6 </td> <td>93.0 </td> <td>85.3 </td> </tr> <tr> <td>Knowledge reasoning </td> <td>TriviaQA-Wiki (5-shot) </td> <td>78.5 </td> <td>72.1 </td> <td>79.6 </td> <td>89.7 </td> <td>87.5 </td> </tr> <tr> <td rowspan="4" >Reading comprehension </td> <td>SQuAD (1-shot) </td> <td>76.4 </td> <td>72.2 </td> <td>72.1 </td> <td>85.6 </td> <td>82.6 </td> </tr> <tr> <td>QuAC (1-shot, F1) </td> <td>44.4 </td> <td>39.6 </td> <td>44.9 </td> <td>51.1 </td> <td>49.4 </td> </tr> <tr> <td>BoolQ (0-shot) </td> <td>75.7 </td> <td>65.5 </td> <td>66.9 </td> <td>79.0 </td> <td>73.1 </td> </tr> <tr> <td>DROP (3-shot, F1) </td> <td>58.4 </td> <td>37.9 </td> <td>49.8 </td> <td>79.7 </td> <td>70.2 </td> </tr> </table> ### Instruction tuned models <table> <tr> <td><strong>Benchmark</strong> </td> <td><strong>Llama 3 8B</strong> </td> <td><strong>Llama 2 7B</strong> </td> <td><strong>Llama 2 13B</strong> </td> <td><strong>Llama 3 70B</strong> </td> <td><strong>Llama 2 70B</strong> </td> </tr> <tr> <td>MMLU (5-shot) </td> <td>68.4 </td> <td>34.1 </td> <td>47.8 </td> <td>82.0 </td> <td>52.9 </td> </tr> <tr> <td>GPQA (0-shot) </td> <td>34.2 </td> <td>21.7 </td> <td>22.3 </td> <td>39.5 </td> <td>21.0 </td> </tr> <tr> <td>HumanEval (0-shot) </td> <td>62.2 </td> <td>7.9 </td> <td>14.0 </td> <td>81.7 </td> <td>25.6 </td> </tr> <tr> <td>GSM-8K (8-shot, CoT) </td> <td>79.6 </td> <td>25.7 </td> <td>77.4 </td> <td>93.0 </td> <td>57.5 </td> </tr> <tr> <td>MATH (4-shot, CoT) </td> <td>30.0 </td> <td>3.8 </td> <td>6.7 </td> <td>50.4 </td> <td>11.6 </td> </tr> </table> ### Responsibility & Safety We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community. Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications. Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience. As part of the Llama 3 release, we updated our [Responsible Use Guide](https://llama.meta.com/responsible-use-guide/) to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including [Meta Llama Guard 2](https://llama.meta.com/purple-llama/) and [Code Shield](https://llama.meta.com/purple-llama/) safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a [reference implementation](https://github.com/meta-llama/llama-recipes/tree/main/recipes/responsible_ai) to get you started. #### Llama 3-Instruct As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case. <span style="text-decoration:underline;">Safety</span> For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable. <span style="text-decoration:underline;">Refusals</span> In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2. We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date. #### Responsible release In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision. Misuse If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy/](https://llama.meta.com/llama3/use-policy/). #### Critical risks <span style="text-decoration:underline;">CBRNE</span> (Chemical, Biological, Radiological, Nuclear, and high yield Explosives) We have conducted a two fold assessment of the safety of the model in this area: * Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks. * Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model). ### <span style="text-decoration:underline;">Cyber Security </span> We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of [equivalent coding capability](https://huggingface.co/spaces/facebook/CyberSecEval). ### <span style="text-decoration:underline;">Child Safety</span> Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences. ### Community Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our [Github repository](https://github.com/meta-llama/PurpleLlama). Finally, we put in place a set of resources including an [output reporting mechanism](https://developers.facebook.com/llama_output_feedback) and [bug bounty program](https://www.facebook.com/whitehat) to continuously improve the Llama technology with the help of the community. ## Ethical Considerations and Limitations The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress. But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating [Purple Llama](https://github.com/facebookresearch/PurpleLlama) solutions into your workflows and specifically [Llama Guard](https://ai.meta.com/research/publications/llama-guard-llm-based-input-output-safeguard-for-human-ai-conversations/) which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety. Please see the Responsible Use Guide available at [http://llama.meta.com/responsible-use-guide](http://llama.meta.com/responsible-use-guide) ## Citation instructions @article{llama3modelcard, title={Llama 3 Model Card}, author={AI@Meta}, year={2024}, url = {https://github.com/meta-llama/llama3/blob/main/MODEL_CARD.md} } ## Contributors Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos	{"language": ["en"], "license": "other", "tags": ["facebook", "meta", "pytorch", "llama", "llama-3"], "pipeline_tag": "text-generation", "license_name": "llama3", "license_link": "LICENSE", "extra_gated_prompt": "### META LLAMA 3 COMMUNITY LICENSE AGREEMENT\nMeta Llama 3 Version Release Date: April 18, 2024\n\"Agreement\" means the terms and conditions for use, reproduction, distribution and modification of the Llama Materials set forth herein.\n\"Documentation\" means the specifications, manuals and documentation accompanying Meta Llama 3 distributed by Meta at https://llama.meta.com/get-started/.\n\"Licensee\" or \"you\" means you, or your employer or any other person or entity (if you are entering into this Agreement on such person or entity\u2019s behalf), of the age required under applicable laws, rules or regulations to provide legal consent and that has legal authority to bind your employer or such other person or entity if you are entering in this Agreement on their behalf.\n\"Meta Llama 3\" means the foundational large language models and software and algorithms, including machine-learning model code, trained model weights, inference-enabling code, training-enabling code, fine-tuning enabling code and other elements of the foregoing distributed by Meta at https://llama.meta.com/llama-downloads.\n\"Llama Materials\" means, collectively, Meta\u2019s proprietary Meta Llama 3 and Documentation (and any portion thereof) made available under this Agreement.\n\"Meta\" or \"we\" means Meta Platforms Ireland Limited (if you are located in or, if you are an entity, your principal place of business is in the EEA or Switzerland) and Meta Platforms, Inc. (if you are located outside of the EEA or Switzerland).\n \n1. License Rights and Redistribution.\na. Grant of Rights. You are granted a non-exclusive, worldwide, non-transferable and royalty-free limited license under Meta\u2019s intellectual property or other rights owned by Meta embodied in the Llama Materials to use, reproduce, distribute, copy, create derivative works of, and make modifications to the Llama Materials.\nb. Redistribution and Use.\ni. If you distribute or make available the Llama Materials (or any derivative works thereof), or a product or service that uses any of them, including another AI model, you shall (A) provide a copy of this Agreement with any such Llama Materials; and (B) prominently display \u201cBuilt with Meta Llama 3\u201d on a related website, user interface, blogpost, about page, or product documentation. If you use the Llama Materials to create, train, fine tune, or otherwise improve an AI model, which is distributed or made available, you shall also include \u201cLlama 3\u201d at the beginning of any such AI model name.\nii. If you receive Llama Materials, or any derivative works thereof, from a Licensee as part of an integrated end user product, then Section 2 of this Agreement will not apply to you.\niii. You must retain in all copies of the Llama Materials that you distribute the following attribution notice within a \u201cNotice\u201d text file distributed as a part of such copies: \u201cMeta Llama 3 is licensed under the Meta Llama 3 Community License, Copyright \u00a9 Meta Platforms, Inc. All Rights Reserved.\u201d\niv. Your use of the Llama Materials must comply with applicable laws and regulations (including trade compliance laws and regulations) and adhere to the Acceptable Use Policy for the Llama Materials (available at https://llama.meta.com/llama3/use-policy), which is hereby incorporated by reference into this Agreement.\nv. You will not use the Llama Materials or any output or results of the Llama Materials to improve any other large language model (excluding Meta Llama 3 or derivative works thereof).\n2. Additional Commercial Terms. If, on the Meta Llama 3 version release date, the monthly active users of the products or services made available by or for Licensee, or Licensee\u2019s affiliates, is greater than 700 million monthly active users in the preceding calendar month, you must request a license from Meta, which Meta may grant to you in its sole discretion, and you are not authorized to exercise any of the rights under this Agreement unless or until Meta otherwise expressly grants you such rights.\n3. Disclaimer of Warranty. UNLESS REQUIRED BY APPLICABLE LAW, THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS THEREFROM ARE PROVIDED ON AN \u201cAS IS\u201d BASIS, WITHOUT WARRANTIES OF ANY KIND, AND META DISCLAIMS ALL WARRANTIES OF ANY KIND, BOTH EXPRESS AND IMPLIED, INCLUDING, WITHOUT LIMITATION, ANY WARRANTIES OF TITLE, NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE. YOU ARE SOLELY RESPONSIBLE FOR DETERMINING THE APPROPRIATENESS OF USING OR REDISTRIBUTING THE LLAMA MATERIALS AND ASSUME ANY RISKS ASSOCIATED WITH YOUR USE OF THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS.\n4. Limitation of Liability. IN NO EVENT WILL META OR ITS AFFILIATES BE LIABLE UNDER ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, TORT, NEGLIGENCE, PRODUCTS LIABILITY, OR OTHERWISE, ARISING OUT OF THIS AGREEMENT, FOR ANY LOST PROFITS OR ANY INDIRECT, SPECIAL, CONSEQUENTIAL, INCIDENTAL, EXEMPLARY OR PUNITIVE DAMAGES, EVEN IF META OR ITS AFFILIATES HAVE BEEN ADVISED OF THE POSSIBILITY OF ANY OF THE FOREGOING.\n5. Intellectual Property.\na. No trademark licenses are granted under this Agreement, and in connection with the Llama Materials, neither Meta nor Licensee may use any name or mark owned by or associated with the other or any of its affiliates, except as required for reasonable and customary use in describing and redistributing the Llama Materials or as set forth in this Section 5(a). Meta hereby grants you a license to use \u201cLlama 3\u201d (the \u201cMark\u201d) solely as required to comply with the last sentence of Section 1.b.i. You will comply with Meta\u2019s brand guidelines (currently accessible at https://about.meta.com/brand/resources/meta/company-brand/ ). All goodwill arising out of your use of the Mark will inure to the benefit of Meta.\nb. Subject to Meta\u2019s ownership of Llama Materials and derivatives made by or for Meta, with respect to any derivative works and modifications of the Llama Materials that are made by you, as between you and Meta, you are and will be the owner of such derivative works and modifications.\nc. If you institute litigation or other proceedings against Meta or any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Llama Materials or Meta Llama 3 outputs or results, or any portion of any of the foregoing, constitutes infringement of intellectual property or other rights owned or licensable by you, then any licenses granted to you under this Agreement shall terminate as of the date such litigation or claim is filed or instituted. You will indemnify and hold harmless Meta from and against any claim by any third party arising out of or related to your use or distribution of the Llama Materials.\n6. Term and Termination. The term of this Agreement will commence upon your acceptance of this Agreement or access to the Llama Materials and will continue in full force and effect until terminated in accordance with the terms and conditions herein. Meta may terminate this Agreement if you are in breach of any term or condition of this Agreement. Upon termination of this Agreement, you shall delete and cease use of the Llama Materials. Sections 3, 4 and 7 shall survive the termination of this Agreement.\n7. Governing Law and Jurisdiction. This Agreement will be governed and construed under the laws of the State of California without regard to choice of law principles, and the UN Convention on Contracts for the International Sale of Goods does not apply to this Agreement. The courts of California shall have exclusive jurisdiction of any dispute arising out of this Agreement.\n### Meta Llama 3 Acceptable Use Policy\nMeta is committed to promoting safe and fair use of its tools and features, including Meta Llama 3. If you access or use Meta Llama 3, you agree to this Acceptable Use Policy (\u201cPolicy\u201d). The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy](https://llama.meta.com/llama3/use-policy)\n#### Prohibited Uses\nWe want everyone to use Meta Llama 3 safely and responsibly. You agree you will not use, or allow others to use, Meta Llama 3 to: 1. Violate the law or others\u2019 rights, including to:\n 1. Engage in, promote, generate, contribute to, encourage, plan, incite, or further illegal or unlawful activity or content, such as:\n 1. Violence or terrorism\n 2. Exploitation or harm to children, including the solicitation, creation, acquisition, or dissemination of child exploitative content or failure to report Child Sexual Abuse Material\n 3. Human trafficking, exploitation, and sexual violence\n 4. The illegal distribution of information or materials to minors, including obscene materials, or failure to employ legally required age-gating in connection with such information or materials.\n 5. Sexual solicitation\n 6. Any other criminal activity\n 2. Engage in, promote, incite, or facilitate the harassment, abuse, threatening, or bullying of individuals or groups of individuals\n 3. Engage in, promote, incite, or facilitate discrimination or other unlawful or harmful conduct in the provision of employment, employment benefits, credit, housing, other economic benefits, or other essential goods and services\n 4. Engage in the unauthorized or unlicensed practice of any profession including, but not limited to, financial, legal, medical/health, or related professional practices\n 5. Collect, process, disclose, generate, or infer health, demographic, or other sensitive personal or private information about individuals without rights and consents required by applicable laws\n 6. Engage in or facilitate any action or generate any content that infringes, misappropriates, or otherwise violates any third-party rights, including the outputs or results of any products or services using the Llama Materials\n 7. Create, generate, or facilitate the creation of malicious code, malware, computer viruses or do anything else that could disable, overburden, interfere with or impair the proper working, integrity, operation or appearance of a website or computer system\n2. Engage in, promote, incite, facilitate, or assist in the planning or development of activities that present a risk of death or bodily harm to individuals, including use of Meta Llama 3 related to the following:\n 1. Military, warfare, nuclear industries or applications, espionage, use for materials or activities that are subject to the International Traffic Arms Regulations (ITAR) maintained by the United States Department of State\n 2. Guns and illegal weapons (including weapon development)\n 3. Illegal drugs and regulated/controlled substances\n 4. Operation of critical infrastructure, transportation technologies, or heavy machinery\n 5. Self-harm or harm to others, including suicide, cutting, and eating disorders\n 6. Any content intended to incite or promote violence, abuse, or any infliction of bodily harm to an individual\n3. Intentionally deceive or mislead others, including use of Meta Llama 3 related to the following:\n 1. Generating, promoting, or furthering fraud or the creation or promotion of disinformation\n 2. Generating, promoting, or furthering defamatory content, including the creation of defamatory statements, images, or other content\n 3. Generating, promoting, or further distributing spam\n 4. Impersonating another individual without consent, authorization, or legal right\n 5. Representing that the use of Meta Llama 3 or outputs are human-generated\n 6. Generating or facilitating false online engagement, including fake reviews and other means of fake online engagement\n4. Fail to appropriately disclose to end users any known dangers of your AI system\nPlease report any violation of this Policy, software \u201cbug,\u201d or other problems that could lead to a violation of this Policy through one of the following means:\n * Reporting issues with the model: [https://github.com/meta-llama/llama3](https://github.com/meta-llama/llama3)\n * Reporting risky content generated by the model:\n developers.facebook.com/llama_output_feedback\n * Reporting bugs and security concerns: facebook.com/whitehat/info\n * Reporting violations of the Acceptable Use Policy or unlicensed uses of Meta Llama 3: [email protected]", "extra_gated_fields": {"First Name": "text", "Last Name": "text", "Date of birth": "date_picker", "Country": "country", "Affiliation": "text", "geo": "ip_location", "By clicking Submit below I accept the terms of the license and acknowledge that the information I provide will be collected stored processed and shared in accordance with the Meta Privacy Policy": "checkbox"}, "extra_gated_description": "The information you provide will be collected, stored, processed and shared in accordance with the [Meta Privacy Policy](https://www.facebook.com/privacy/policy/).", "extra_gated_button_content": "Submit"}	LoneStriker/Meta-Llama-3-70B-GGUF	null	[ "gguf", "facebook", "meta", "pytorch", "llama", "llama-3", "text-generation", "en", "license:other", "region:us" ]	null	2024-04-19T02:51:36+00:00	[]	[ "en" ]	TAGS #gguf #facebook #meta #pytorch #llama #llama-3 #text-generation #en #license-other #region-us	Model Details ------------- Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety. Model developers Meta Variations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants. Input Models input text only. Output Models generate text and code only. Model Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. Llama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability. Model Release Date April 18, 2024. Status This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback. License A custom commercial license is available at: URL Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model README. For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go here. Intended Use ------------ Intended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. Out-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English. Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. How to use ---------- This repository contains two versions of Meta-Llama-3-8B-Instruct, for use with transformers and with the original 'llama3' codebase. ### Use with transformers See the snippet below for usage with Transformers: ### Use with 'llama3' Please, follow the instructions in the repository. To download Original checkpoints, see the example command below leveraging 'huggingface-cli': For Hugging Face support, we recommend using transformers or TGI, but a similar command works. Hardware and Software --------------------- Training Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute. Carbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program. CO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others. Training Data ------------- Overview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data. Data Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively. Benchmarks ---------- In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here. ### Base pretrained models ### Instruction tuned models ### Responsibility & Safety We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community. Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications. Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience. As part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started. #### Llama 3-Instruct As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case. Safety For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable. Refusals In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2. We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date. #### Responsible release In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision. Misuse If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL #### Critical risks CBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives) We have conducted a two fold assessment of the safety of the model in this area: * Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks. * Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model). ### Cyber Security We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability. ### Child Safety Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences. ### Community Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository. Finally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community. Ethical Considerations and Limitations -------------------------------------- The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress. But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety. Please see the Responsible Use Guide available at URL instructions @article{llama3modelcard, title={Llama 3 Model Card}, author={AI@Meta}, year={2024}, url = {URL } Contributors ------------ Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos	[ "### Use with transformers\n\n\nSee the snippet below for usage with Transformers:", "### Use with 'llama3'\n\n\nPlease, follow the instructions in the repository.\n\n\nTo download Original checkpoints, see the example command below leveraging 'huggingface-cli':\n\n\nFor Hugging Face support, we recommend using transformers or TGI, but a similar command works.\n\n\nHardware and Software\n---------------------\n\n\nTraining Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.\n\n\nCarbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.\n\n\n\nCO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.\n\n\nTraining Data\n-------------\n\n\nOverview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.\n\n\nData Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.\n\n\nBenchmarks\n----------\n\n\nIn this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here.", "### Base pretrained models", "### Instruction tuned models", "### Responsibility & Safety\n\n\nWe believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.\n\n\nFoundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.\n\n\nRather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.\n\n\nAs part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started.", "#### Llama 3-Instruct\n\n\nAs outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.\n\n\nSafety\n\n\nFor our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.\n\n\nRefusals\n\n\nIn addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.\n\n\nWe built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.", "#### Responsible release\n\n\nIn addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.\n\n\nMisuse\n\n\nIf you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL", "#### Critical risks\n\n\nCBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)\n\n\nWe have conducted a two fold assessment of the safety of the model in this area:\n\n\n* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.\n* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).", "### Cyber Security\n\n\nWe have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability.", "### Child Safety\n\n\nChild Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.", "### Community\n\n\nGenerative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository.\n\n\nFinally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community.\n\n\nEthical Considerations and Limitations\n--------------------------------------\n\n\nThe core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.\n\n\nBut Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.\n\n\nPlease see the Responsible Use Guide available at URL\n\n\ninstructions\n\n\n@article{llama3modelcard,\n\n\ntitle={Llama 3 Model Card},\n\n\nauthor={AI@Meta},\n\n\nyear={2024},\n\n\nurl = {URL\n\n\n}\n\n\nContributors\n------------\n\n\nAaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos" ]	[ "TAGS\n#gguf #facebook #meta #pytorch #llama #llama-3 #text-generation #en #license-other #region-us \n", "### Use with transformers\n\n\nSee the snippet below for usage with Transformers:", "### Use with 'llama3'\n\n\nPlease, follow the instructions in the repository.\n\n\nTo download Original checkpoints, see the example command below leveraging 'huggingface-cli':\n\n\nFor Hugging Face support, we recommend using transformers or TGI, but a similar command works.\n\n\nHardware and Software\n---------------------\n\n\nTraining Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.\n\n\nCarbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.\n\n\n\nCO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.\n\n\nTraining Data\n-------------\n\n\nOverview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.\n\n\nData Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.\n\n\nBenchmarks\n----------\n\n\nIn this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here.", "### Base pretrained models", "### Instruction tuned models", "### Responsibility & Safety\n\n\nWe believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.\n\n\nFoundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.\n\n\nRather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.\n\n\nAs part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started.", "#### Llama 3-Instruct\n\n\nAs outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.\n\n\nSafety\n\n\nFor our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.\n\n\nRefusals\n\n\nIn addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.\n\n\nWe built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.", "#### Responsible release\n\n\nIn addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.\n\n\nMisuse\n\n\nIf you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL", "#### Critical risks\n\n\nCBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)\n\n\nWe have conducted a two fold assessment of the safety of the model in this area:\n\n\n* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.\n* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).", "### Cyber Security\n\n\nWe have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability.", "### Child Safety\n\n\nChild Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.", "### Community\n\n\nGenerative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository.\n\n\nFinally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community.\n\n\nEthical Considerations and Limitations\n--------------------------------------\n\n\nThe core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.\n\n\nBut Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.\n\n\nPlease see the Responsible Use Guide available at URL\n\n\ninstructions\n\n\n@article{llama3modelcard,\n\n\ntitle={Llama 3 Model Card},\n\n\nauthor={AI@Meta},\n\n\nyear={2024},\n\n\nurl = {URL\n\n\n}\n\n\nContributors\n------------\n\n\nAaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos" ]
text-classification	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # augmented_distillbert This model is a fine-tuned version of [distilbert/distilbert-base-uncased](https://huggingface.co/distilbert/distilbert-base-uncased) on the None dataset. It achieves the following results on the evaluation set: - Loss: 0.4493 - Accuracy: 0.9195 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 4 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - lr_scheduler_warmup_ratio: 0.01 - num_epochs: 2 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\| \| 0.0002 \| 0.2 \| 558 \| 0.8604 \| 0.8762 \| \| 0.002 \| 0.4 \| 1116 \| 0.4315 \| 0.9057 \| \| 1.4422 \| 0.6 \| 1674 \| 0.3956 \| 0.9119 \| \| 0.2888 \| 0.8 \| 2232 \| 0.3751 \| 0.9110 \| \| 0.0061 \| 1.0 \| 2790 \| 0.3465 \| 0.9181 \| \| 0.0007 \| 1.2 \| 3348 \| 0.4745 \| 0.9110 \| \| 0.0018 \| 1.4 \| 3906 \| 0.4281 \| 0.9238 \| \| 0.0005 \| 1.6 \| 4464 \| 0.4390 \| 0.92 \| \| 0.0028 \| 1.8 \| 5022 \| 0.4493 \| 0.9195 \| ### Framework versions - Transformers 4.38.2 - Pytorch 2.2.1+cu121 - Datasets 2.18.0 - Tokenizers 0.15.2	{"license": "apache-2.0", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "distilbert/distilbert-base-uncased", "model-index": [{"name": "augmented_distillbert", "results": []}]}	erostrate9/augmented_distillbert	null	[ "transformers", "safetensors", "distilbert", "text-classification", "generated_from_trainer", "base_model:distilbert/distilbert-base-uncased", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T02:53:19+00:00	[]	[]	TAGS #transformers #safetensors #distilbert #text-classification #generated_from_trainer #base_model-distilbert/distilbert-base-uncased #license-apache-2.0 #autotrain_compatible #endpoints_compatible #region-us	augmented\_distillbert ====================== This model is a fine-tuned version of distilbert/distilbert-base-uncased on the None dataset. It achieves the following results on the evaluation set: * Loss: 0.4493 * Accuracy: 0.9195 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * learning\_rate: 5e-05 * train\_batch\_size: 4 * eval\_batch\_size: 16 * seed: 42 * optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 * lr\_scheduler\_type: cosine * lr\_scheduler\_warmup\_ratio: 0.01 * num\_epochs: 2 ### Training results ### Framework versions * Transformers 4.38.2 * Pytorch 2.2.1+cu121 * Datasets 2.18.0 * Tokenizers 0.15.2	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 5e-05\n* train\\_batch\\_size: 4\n* eval\\_batch\\_size: 16\n* seed: 42\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: cosine\n* lr\\_scheduler\\_warmup\\_ratio: 0.01\n* num\\_epochs: 2", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* Pytorch 2.2.1+cu121\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #safetensors #distilbert #text-classification #generated_from_trainer #base_model-distilbert/distilbert-base-uncased #license-apache-2.0 #autotrain_compatible #endpoints_compatible #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 5e-05\n* train\\_batch\\_size: 4\n* eval\\_batch\\_size: 16\n* seed: 42\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: cosine\n* lr\\_scheduler\\_warmup\\_ratio: 0.01\n* num\\_epochs: 2", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* Pytorch 2.2.1+cu121\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]
text-generation	null	# Quant Infos ## Includes latest bpe tokenizer fixes 🎉 - Updated for latest bpe pre-tokenizer fixes https://github.com/ggerganov/llama.cpp/pull/6920 - quants done with an importance matrix for improved quantization loss - K & IQ quants in basically all variants from Q6_K down to IQ1_S - fixed end token for instruct mode (<\|eot_id\|>[128009]) - Quantized with [llama.cpp](https://github.com/ggerganov/llama.cpp) commit [f4ab2a41476600a98067a9474ea8f9e6db41bcfa](https://github.com/ggerganov/llama.cpp/commit/f4ab2a41476600a98067a9474ea8f9e6db41bcfa) (master from 2024-04-29) - Imatrtix generated with [this](https://github.com/ggerganov/llama.cpp/discussions/5263#discussioncomment-8395384) dataset. ``` ./imatrix -c 512 -m $model_name-f16.gguf -f $llama_cpp_path/groups_merged.txt -o $out_path/imat-f16-gmerged.dat ``` ## Note about recent tokenizer fixes The newest quants uploaded here need at least commit f4ab2a41476600a98067a9474ea8f9e6db41bcfa, this is not integrated into most upstream tools yet as it was just released. (29-04-24) ## Note about eos token Llama 3 uses a different eos tokens depending if it is in instruct mode. The initial upload had some issues with this as it uses the "default" eos token of 128001, but when in instruct mode llama only outputs 128009 as eos token which causes it to ramble on and on without stopping. I have uploaded fixed quants with the eos token id manually set to 128009. This fixes the issue for me, but you have to make sure to use the correct chat template, ~I recommend using [this](https://github.com/ggerganov/llama.cpp/pull/6751) PR~(it has been merged just used the newest llama.cpp master) and then launching llama.cpp with `--chat-template llama3`. If you do not want to redownload you can fix your local gguf file with this command: ``` python3 ./path-to-llama.cpp/gguf-py/scripts/gguf-set-metadata.py $file tokenizer.ggml.eos_token_id 128009 --force ``` ------------------------ # Original Model Card: ## Model Details Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety. Model developers Meta Variations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants. Input Models input text only. Output Models generate text and code only. Model Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. <table> <tr> <td> </td> <td><strong>Training Data</strong> </td> <td><strong>Params</strong> </td> <td><strong>Context length</strong> </td> <td><strong>GQA</strong> </td> <td><strong>Token count</strong> </td> <td><strong>Knowledge cutoff</strong> </td> </tr> <tr> <td rowspan="2" >Llama 3 </td> <td rowspan="2" >A new mix of publicly available online data. </td> <td>8B </td> <td>8k </td> <td>Yes </td> <td rowspan="2" >15T+ </td> <td>March, 2023 </td> </tr> <tr> <td>70B </td> <td>8k </td> <td>Yes </td> <td>December, 2023 </td> </tr> </table> Llama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability. Model Release Date April 18, 2024. Status This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback. License A custom commercial license is available at: [https://llama.meta.com/llama3/license](https://llama.meta.com/llama3/license) Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model [README](https://github.com/meta-llama/llama3). For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go [here](https://github.com/meta-llama/llama-recipes). ## Intended Use Intended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. Out-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English. Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. ## How to use This repository contains two versions of Meta-Llama-3-8B-Instruct, for use with transformers and with the original `llama3` codebase. ### Use with transformers See the snippet below for usage with Transformers: ```python import transformers import torch model_id = "meta-llama/Meta-Llama-3-8B-Instruct" pipeline = transformers.pipeline( "text-generation", model=model_id, model_kwargs={"torch_dtype": torch.bfloat16}, device="auto", ) messages = [ {"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"}, {"role": "user", "content": "Who are you?"}, ] prompt = pipeline.tokenizer.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) terminators = [ pipeline.tokenizer.eos_token_id, pipeline.tokenizer.convert_tokens_to_ids("<\|eot_id\|>") ] outputs = pipeline( prompt, max_new_tokens=256, eos_token_id=terminators, do_sample=True, temperature=0.6, top_p=0.9, ) print(outputs[0]["generated_text"][len(prompt):]) ``` ### Use with `llama3` Please, follow the instructions in the [repository](https://github.com/meta-llama/llama3) To download Original checkpoints, see the example command below leveraging `huggingface-cli`: ``` huggingface-cli download meta-llama/Meta-Llama-3-8B-Instruct --include "original/" --local-dir Meta-Llama-3-8B-Instruct ``` For Hugging Face support, we recommend using transformers or TGI, but a similar command works. ## Hardware and Software Training Factors* We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute. Carbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program. <table> <tr> <td> </td> <td><strong>Time (GPU hours)</strong> </td> <td><strong>Power Consumption (W)</strong> </td> <td><strong>Carbon Emitted(tCO2eq)</strong> </td> </tr> <tr> <td>Llama 3 8B </td> <td>1.3M </td> <td>700 </td> <td>390 </td> </tr> <tr> <td>Llama 3 70B </td> <td>6.4M </td> <td>700 </td> <td>1900 </td> </tr> <tr> <td>Total </td> <td>7.7M </td> <td> </td> <td>2290 </td> </tr> </table> CO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others. ## Training Data Overview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data. Data Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively. ## Benchmarks In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see [here](https://github.com/meta-llama/llama3/blob/main/eval_methodology.md). ### Base pretrained models <table> <tr> <td><strong>Category</strong> </td> <td><strong>Benchmark</strong> </td> <td><strong>Llama 3 8B</strong> </td> <td><strong>Llama2 7B</strong> </td> <td><strong>Llama2 13B</strong> </td> <td><strong>Llama 3 70B</strong> </td> <td><strong>Llama2 70B</strong> </td> </tr> <tr> <td rowspan="6" >General </td> <td>MMLU (5-shot) </td> <td>66.6 </td> <td>45.7 </td> <td>53.8 </td> <td>79.5 </td> <td>69.7 </td> </tr> <tr> <td>AGIEval English (3-5 shot) </td> <td>45.9 </td> <td>28.8 </td> <td>38.7 </td> <td>63.0 </td> <td>54.8 </td> </tr> <tr> <td>CommonSenseQA (7-shot) </td> <td>72.6 </td> <td>57.6 </td> <td>67.6 </td> <td>83.8 </td> <td>78.7 </td> </tr> <tr> <td>Winogrande (5-shot) </td> <td>76.1 </td> <td>73.3 </td> <td>75.4 </td> <td>83.1 </td> <td>81.8 </td> </tr> <tr> <td>BIG-Bench Hard (3-shot, CoT) </td> <td>61.1 </td> <td>38.1 </td> <td>47.0 </td> <td>81.3 </td> <td>65.7 </td> </tr> <tr> <td>ARC-Challenge (25-shot) </td> <td>78.6 </td> <td>53.7 </td> <td>67.6 </td> <td>93.0 </td> <td>85.3 </td> </tr> <tr> <td>Knowledge reasoning </td> <td>TriviaQA-Wiki (5-shot) </td> <td>78.5 </td> <td>72.1 </td> <td>79.6 </td> <td>89.7 </td> <td>87.5 </td> </tr> <tr> <td rowspan="4" >Reading comprehension </td> <td>SQuAD (1-shot) </td> <td>76.4 </td> <td>72.2 </td> <td>72.1 </td> <td>85.6 </td> <td>82.6 </td> </tr> <tr> <td>QuAC (1-shot, F1) </td> <td>44.4 </td> <td>39.6 </td> <td>44.9 </td> <td>51.1 </td> <td>49.4 </td> </tr> <tr> <td>BoolQ (0-shot) </td> <td>75.7 </td> <td>65.5 </td> <td>66.9 </td> <td>79.0 </td> <td>73.1 </td> </tr> <tr> <td>DROP (3-shot, F1) </td> <td>58.4 </td> <td>37.9 </td> <td>49.8 </td> <td>79.7 </td> <td>70.2 </td> </tr> </table> ### Instruction tuned models <table> <tr> <td><strong>Benchmark</strong> </td> <td><strong>Llama 3 8B</strong> </td> <td><strong>Llama 2 7B</strong> </td> <td><strong>Llama 2 13B</strong> </td> <td><strong>Llama 3 70B</strong> </td> <td><strong>Llama 2 70B</strong> </td> </tr> <tr> <td>MMLU (5-shot) </td> <td>68.4 </td> <td>34.1 </td> <td>47.8 </td> <td>82.0 </td> <td>52.9 </td> </tr> <tr> <td>GPQA (0-shot) </td> <td>34.2 </td> <td>21.7 </td> <td>22.3 </td> <td>39.5 </td> <td>21.0 </td> </tr> <tr> <td>HumanEval (0-shot) </td> <td>62.2 </td> <td>7.9 </td> <td>14.0 </td> <td>81.7 </td> <td>25.6 </td> </tr> <tr> <td>GSM-8K (8-shot, CoT) </td> <td>79.6 </td> <td>25.7 </td> <td>77.4 </td> <td>93.0 </td> <td>57.5 </td> </tr> <tr> <td>MATH (4-shot, CoT) </td> <td>30.0 </td> <td>3.8 </td> <td>6.7 </td> <td>50.4 </td> <td>11.6 </td> </tr> </table> ### Responsibility & Safety We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community. Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications. Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience. As part of the Llama 3 release, we updated our [Responsible Use Guide](https://llama.meta.com/responsible-use-guide/) to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including [Meta Llama Guard 2](https://llama.meta.com/purple-llama/) and [Code Shield](https://llama.meta.com/purple-llama/) safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a [reference implementation](https://github.com/meta-llama/llama-recipes/tree/main/recipes/responsible_ai) to get you started. #### Llama 3-Instruct As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case. <span style="text-decoration:underline;">Safety</span> For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable. <span style="text-decoration:underline;">Refusals</span> In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2. We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date. #### Responsible release In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision. Misuse If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy/](https://llama.meta.com/llama3/use-policy/). #### Critical risks <span style="text-decoration:underline;">CBRNE</span> (Chemical, Biological, Radiological, Nuclear, and high yield Explosives) We have conducted a two fold assessment of the safety of the model in this area: * Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks. * Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model). ### <span style="text-decoration:underline;">Cyber Security </span> We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of [equivalent coding capability](https://huggingface.co/spaces/facebook/CyberSecEval). ### <span style="text-decoration:underline;">Child Safety</span> Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences. ### Community Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our [Github repository](https://github.com/meta-llama/PurpleLlama). Finally, we put in place a set of resources including an [output reporting mechanism](https://developers.facebook.com/llama_output_feedback) and [bug bounty program](https://www.facebook.com/whitehat) to continuously improve the Llama technology with the help of the community. ## Ethical Considerations and Limitations The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress. But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating [Purple Llama](https://github.com/facebookresearch/PurpleLlama) solutions into your workflows and specifically [Llama Guard](https://ai.meta.com/research/publications/llama-guard-llm-based-input-output-safeguard-for-human-ai-conversations/) which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety. Please see the Responsible Use Guide available at [http://llama.meta.com/responsible-use-guide](http://llama.meta.com/responsible-use-guide) ## Citation instructions @article{llama3modelcard, title={Llama 3 Model Card}, author={AI@Meta}, year={2024}, url = {https://github.com/meta-llama/llama3/blob/main/MODEL_CARD.md} } ## Contributors Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos	{"language": ["en"], "license": "other", "tags": ["facebook", "meta", "pytorch", "llama", "llama-3", "llama.cpp", "gguf", "imatrix", "importance matrix"], "base_model": "meta-llama/Meta-Llama-3-8B-Instruct", "pipeline_tag": "text-generation", "license_name": "llama3", "license_link": "LICENSE"}	qwp4w3hyb/Meta-Llama-3-8B-Instruct-iMat-GGUF	null	[ "gguf", "facebook", "meta", "pytorch", "llama", "llama-3", "llama.cpp", "imatrix", "importance matrix", "text-generation", "en", "base_model:meta-llama/Meta-Llama-3-8B-Instruct", "license:other", "region:us" ]	null	2024-04-19T02:56:54+00:00	[]	[ "en" ]	TAGS #gguf #facebook #meta #pytorch #llama #llama-3 #llama.cpp #imatrix #importance matrix #text-generation #en #base_model-meta-llama/Meta-Llama-3-8B-Instruct #license-other #region-us	Quant Infos =========== Includes latest bpe tokenizer fixes ----------------------------------- * Updated for latest bpe pre-tokenizer fixes URL * quants done with an importance matrix for improved quantization loss * K & IQ quants in basically all variants from Q6\_K down to IQ1\_S * fixed end token for instruct mode (<\|eot\_id\|>[128009]) * Quantized with URL commit f4ab2a41476600a98067a9474ea8f9e6db41bcfa (master from 2024-04-29) * Imatrtix generated with this dataset. Note about recent tokenizer fixes --------------------------------- The newest quants uploaded here need at least commit f4ab2a41476600a98067a9474ea8f9e6db41bcfa, this is not integrated into most upstream tools yet as it was just released. (29-04-24) Note about eos token -------------------- Llama 3 uses a different eos tokens depending if it is in instruct mode. The initial upload had some issues with this as it uses the "default" eos token of 128001, but when in instruct mode llama only outputs 128009 as eos token which causes it to ramble on and on without stopping. I have uploaded fixed quants with the eos token id manually set to 128009. This fixes the issue for me, but you have to make sure to use the correct chat template, ~I recommend using this PR~(it has been merged just used the newest URL master) and then launching URL with '--chat-template llama3'. If you do not want to redownload you can fix your local gguf file with this command: --- Original Model Card: ==================== Model Details ------------- Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety. Model developers Meta Variations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants. Input Models input text only. Output Models generate text and code only. Model Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. Llama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability. Model Release Date April 18, 2024. Status This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback. License A custom commercial license is available at: URL Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model README. For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go here. Intended Use ------------ Intended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. Out-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English. Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. How to use ---------- This repository contains two versions of Meta-Llama-3-8B-Instruct, for use with transformers and with the original 'llama3' codebase. ### Use with transformers See the snippet below for usage with Transformers: ### Use with 'llama3' Please, follow the instructions in the repository To download Original checkpoints, see the example command below leveraging 'huggingface-cli': For Hugging Face support, we recommend using transformers or TGI, but a similar command works. Hardware and Software --------------------- Training Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute. Carbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program. CO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others. Training Data ------------- Overview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data. Data Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively. Benchmarks ---------- In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here. ### Base pretrained models ### Instruction tuned models ### Responsibility & Safety We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community. Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications. Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience. As part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started. #### Llama 3-Instruct As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case. Safety For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable. Refusals In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2. We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date. #### Responsible release In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision. Misuse If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL #### Critical risks CBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives) We have conducted a two fold assessment of the safety of the model in this area: * Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks. * Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model). ### Cyber Security We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability. ### Child Safety Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences. ### Community Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository. Finally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community. Ethical Considerations and Limitations -------------------------------------- The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress. But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety. Please see the Responsible Use Guide available at URL instructions @article{llama3modelcard, title={Llama 3 Model Card}, author={AI@Meta}, year={2024}, url = {URL } Contributors ------------ Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos	[ "### Use with transformers\n\n\nSee the snippet below for usage with Transformers:", "### Use with 'llama3'\n\n\nPlease, follow the instructions in the repository\n\n\nTo download Original checkpoints, see the example command below leveraging 'huggingface-cli':\n\n\nFor Hugging Face support, we recommend using transformers or TGI, but a similar command works.\n\n\nHardware and Software\n---------------------\n\n\nTraining Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.\n\n\nCarbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.\n\n\n\nCO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.\n\n\nTraining Data\n-------------\n\n\nOverview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.\n\n\nData Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.\n\n\nBenchmarks\n----------\n\n\nIn this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here.", "### Base pretrained models", "### Instruction tuned models", "### Responsibility & Safety\n\n\nWe believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.\n\n\nFoundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.\n\n\nRather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.\n\n\nAs part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started.", "#### Llama 3-Instruct\n\n\nAs outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.\n\n\nSafety\n\n\nFor our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.\n\n\nRefusals\n\n\nIn addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.\n\n\nWe built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.", "#### Responsible release\n\n\nIn addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.\n\n\nMisuse\n\n\nIf you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL", "#### Critical risks\n\n\nCBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)\n\n\nWe have conducted a two fold assessment of the safety of the model in this area:\n\n\n* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.\n* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).", "### Cyber Security\n\n\nWe have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability.", "### Child Safety\n\n\nChild Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.", "### Community\n\n\nGenerative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository.\n\n\nFinally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community.\n\n\nEthical Considerations and Limitations\n--------------------------------------\n\n\nThe core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.\n\n\nBut Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.\n\n\nPlease see the Responsible Use Guide available at URL\n\n\ninstructions\n\n\n@article{llama3modelcard,\n\n\ntitle={Llama 3 Model Card},\n\n\nauthor={AI@Meta},\n\n\nyear={2024},\n\n\nurl = {URL\n\n\n}\n\n\nContributors\n------------\n\n\nAaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos" ]	[ "TAGS\n#gguf #facebook #meta #pytorch #llama #llama-3 #llama.cpp #imatrix #importance matrix #text-generation #en #base_model-meta-llama/Meta-Llama-3-8B-Instruct #license-other #region-us \n", "### Use with transformers\n\n\nSee the snippet below for usage with Transformers:", "### Use with 'llama3'\n\n\nPlease, follow the instructions in the repository\n\n\nTo download Original checkpoints, see the example command below leveraging 'huggingface-cli':\n\n\nFor Hugging Face support, we recommend using transformers or TGI, but a similar command works.\n\n\nHardware and Software\n---------------------\n\n\nTraining Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.\n\n\nCarbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.\n\n\n\nCO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.\n\n\nTraining Data\n-------------\n\n\nOverview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.\n\n\nData Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.\n\n\nBenchmarks\n----------\n\n\nIn this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here.", "### Base pretrained models", "### Instruction tuned models", "### Responsibility & Safety\n\n\nWe believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.\n\n\nFoundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.\n\n\nRather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.\n\n\nAs part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started.", "#### Llama 3-Instruct\n\n\nAs outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.\n\n\nSafety\n\n\nFor our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.\n\n\nRefusals\n\n\nIn addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.\n\n\nWe built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.", "#### Responsible release\n\n\nIn addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.\n\n\nMisuse\n\n\nIf you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL", "#### Critical risks\n\n\nCBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)\n\n\nWe have conducted a two fold assessment of the safety of the model in this area:\n\n\n* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.\n* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).", "### Cyber Security\n\n\nWe have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability.", "### Child Safety\n\n\nChild Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.", "### Community\n\n\nGenerative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository.\n\n\nFinally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community.\n\n\nEthical Considerations and Limitations\n--------------------------------------\n\n\nThe core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.\n\n\nBut Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.\n\n\nPlease see the Responsible Use Guide available at URL\n\n\ninstructions\n\n\n@article{llama3modelcard,\n\n\ntitle={Llama 3 Model Card},\n\n\nauthor={AI@Meta},\n\n\nyear={2024},\n\n\nurl = {URL\n\n\n}\n\n\nContributors\n------------\n\n\nAaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos" ]
text-generation	transformers	# NikolayKozloff/fr.brain.carotte-7B-Q8_0-GGUF This model was converted to GGUF format from [`lbl/fr.brain.carotte-7B`](https://huggingface.co/lbl/fr.brain.carotte-7B) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space. Refer to the [original model card](https://huggingface.co/lbl/fr.brain.carotte-7B) for more details on the model. ## Use with llama.cpp Install llama.cpp through brew. ```bash brew install ggerganov/ggerganov/llama.cpp ``` Invoke the llama.cpp server or the CLI. CLI: ```bash llama-cli --hf-repo NikolayKozloff/fr.brain.carotte-7B-Q8_0-GGUF --model fr.brain.carotte-7b.Q8_0.gguf -p "The meaning to life and the universe is" ``` Server: ```bash llama-server --hf-repo NikolayKozloff/fr.brain.carotte-7B-Q8_0-GGUF --model fr.brain.carotte-7b.Q8_0.gguf -c 2048 ``` Note: You can also use this checkpoint directly through the [usage steps](https://github.com/ggerganov/llama.cpp?tab=readme-ov-file#usage) listed in the Llama.cpp repo as well. ``` git clone https://github.com/ggerganov/llama.cpp && cd llama.cpp && make && ./main -m fr.brain.carotte-7b.Q8_0.gguf -n 128 ```	{"language": ["fr", "en"], "license": "apache-2.0", "library_name": "transformers", "tags": ["llama-cpp", "gguf-my-repo"], "pipeline_tag": "text-generation"}	NikolayKozloff/fr.brain.carotte-7B-Q8_0-GGUF	null	[ "transformers", "gguf", "llama-cpp", "gguf-my-repo", "text-generation", "fr", "en", "license:apache-2.0", "endpoints_compatible", "region:us" ]	null	2024-04-19T02:58:52+00:00	[]	[ "fr", "en" ]	TAGS #transformers #gguf #llama-cpp #gguf-my-repo #text-generation #fr #en #license-apache-2.0 #endpoints_compatible #region-us	# NikolayKozloff/URL.carotte-7B-Q8_0-GGUF This model was converted to GGUF format from 'lbl/URL.carotte-7B' using URL via the URL's GGUF-my-repo space. Refer to the original model card for more details on the model. ## Use with URL Install URL through brew. Invoke the URL server or the CLI. CLI: Server: Note: You can also use this checkpoint directly through the usage steps listed in the URL repo as well.	[ "# NikolayKozloff/URL.carotte-7B-Q8_0-GGUF\nThis model was converted to GGUF format from 'lbl/URL.carotte-7B' using URL via the URL's GGUF-my-repo space.\nRefer to the original model card for more details on the model.", "## Use with URL\n\nInstall URL through brew.\n\n\nInvoke the URL server or the CLI.\n\nCLI:\n\n\n\nServer:\n\n\n\nNote: You can also use this checkpoint directly through the usage steps listed in the URL repo as well." ]	[ "TAGS\n#transformers #gguf #llama-cpp #gguf-my-repo #text-generation #fr #en #license-apache-2.0 #endpoints_compatible #region-us \n", "# NikolayKozloff/URL.carotte-7B-Q8_0-GGUF\nThis model was converted to GGUF format from 'lbl/URL.carotte-7B' using URL via the URL's GGUF-my-repo space.\nRefer to the original model card for more details on the model.", "## Use with URL\n\nInstall URL through brew.\n\n\nInvoke the URL server or the CLI.\n\nCLI:\n\n\n\nServer:\n\n\n\nNote: You can also use this checkpoint directly through the usage steps listed in the URL repo as well." ]
text-generation	transformers	# Model Card for Merlinite 7b 🔥 [Paper](https://arxiv.org/abs/2403.01081) ### Overview ![Screenshot 2024-02-22 at 11.26.13 AM.png](model-card/Model%20Card%20for%20Merlinite%207b%2028cc0b72cf574a4a828140d3539ede4a_Screenshot_2024-02-22_at_11.26.13_AM.png) ### Performance \| Model \| Alignment \| Base \| Teacher \| MTBench (Avg) * \| MMLU(5-shot) \| ARC-C(25-shot) \| HellaSwag(10-shot) \| Winogrande(5-shot) \| GSM8K(5-shot- strict) \| \| --- \| --- \| --- \| --- \| --- \| --- \| --- \| --- \| --- \| --- \| \| [Llama-2-13b-chat-hf](https://huggingface.co/meta-llama/Llama-2-13b-chat-hf) \| RLHF \| Llama-2-13b \| Human Annotators \| 6.65 \| 54.58 \| 59.81 \| 82.52 \| 75.93 \| 34.80 \| \| [Orca-2-13b](https://huggingface.co/microsoft/Orca-2-13b) \| Progressive Training \| Llama-2-13b \| GPT-4 \| 6.15 \| 60.37 * \| 59.73 \| 79.86 \| 78.22 \| 48.22 \| \| [WizardLM-13B-V1.2](https://huggingface.co/WizardLM/WizardLM-13B-V1.2) \| Evol-Instruct \| Llama-2-13b \| GPT-4 \| 7.20 \| 54.83 \| 60.24 \| 82.62 \| 76.40 \| 43.75 \| \| [Labradorite-13b](https://huggingface.co/ibm/labradorite-13b) \| Large-scale Alignment for chatBots (LAB) \| Llama-2-13b \| Mixtral-8x7B-Instruct \| 7.23 \| 58.89 \| 61.69 \| 83.15 \| 79.56 \| 40.11 \| \| [Mistral-7B-Instruct-v0.1](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.1) \| SFT \| Mistral-7B-v0.1 \| - \| 6.84 \| 60.37 \| 63.65 \| 84.76 \| 76.80 \| 41.85 \| \| [zephyr-7b-beta](https://huggingface.co/HuggingFaceH4/zephyr-7b-beta) \| SFT/DPO \| Mistral-7B-v0.1 \| GPT-4 \| 7.34 \| 61.07 \| 63.74 \| 84.19 \| 78.06 \| 34.04 \| \| [Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2) \| SFT \| Mistral-7B-v0.1 \| - \| 7.6** \| 60.78 \| 63.14 \| 84.88 \| 77.19 \| 40.03 \| \| Merlinite-7b \| Large-scale Alignment for chatBots (LAB) \| Mistral-7B-v0.1 \| Mixtral-8x7B-Instruct \| 7.66 \| 64.88 \| 63.99 \| 84.37 \| 78.24 \| 44.58 \| [] Numbers for models other than Merlinite-7b and [Labradorite-13b](https://huggingface.co/ibm/labradorite-13b) (ours) are taken from [lmsys/chatbot-arena-leaderboard](https://huggingface.co/spaces/lmsys/chatbot-arena-leaderboard) [] Numbers taken from [MistralAI Release Blog](https://mistral.ai/news/la-plateforme/) ### Method LAB: Large-scale Alignment for chatBots is a novel synthetic data-based alignment tuning method for LLMs from IBM Research. Merlinite-7b is a Mistral-7b-derivative model trained with the LAB methodology, using Mixtral-8x7b-Instruct as a teacher model. LAB consists of three key components: 1. Taxonomy-driven data curation process 2. Large-scale synthetic data generator 3. Two-phased-training with replay buffers ![Untitled](model-card/Model%20Card%20for%20Merlinite%207b%2028cc0b72cf574a4a828140d3539ede4a_Untitled.png) LAB approach allows for adding new knowledge and skills, in an incremental fashion, to an already pre-trained model without suffering from catastrophic forgetting. Taxonomy is a tree of seed examples that are used to prompt a teacher model to generate synthetic data. Taxonomy allows the data curator or the model designer to easily specify a diverse set of the knowledge-domains and skills that they would like to include in their LLM. At a high level, these can be categorized into three high-level bins - knowledge, foundational skills, and compositional skills. The leaf nodes of the taxonomy are tasks associated with one or more seed examples. ![Untitled](model-card/Model%20Card%20for%20Merlinite%207b%2028cc0b72cf574a4a828140d3539ede4a_Untitled%201.png) During the synthetic data generation, unlike previous approaches where seed examples are uniformly drawn from the entire pool (i.e. self-instruct), we use the taxonomy to drive the sampling process: For each knowledge/skill, we only use the local examples within the leaf node as seeds to prompt the teacher model. This makes the teacher model better exploit the task distributions defined by the local examples of each node and the diversity in the taxonomy itself ensures the entire generation covers a wide range of tasks, as illustrated below. In turns, this allows for using Mixtral 8x7B as the teacher model for generation while performing very competitively with models such as ORCA-2, WizardLM, and Zephyr Beta that rely on synthetic data generated by much larger and capable models like GPT-4. ![intuition.png](model-card/Model%20Card%20for%20Merlinite%207b%2028cc0b72cf574a4a828140d3539ede4a_intuition.png) For adding new domain-specific knowledge, we provide an external knowledge source (document) and prompt the model to generate questions and answers based on the document. Foundational skills such as reasoning and compositional skills such as creative writing are generated through in-context learning using the seed examples from the taxonomy. Additionally, to ensure the data is high-quality and safe, we employ steps to check the questions and answers to ensure that they are grounded and safe. This is done using the same teacher model that generated the data. Our training consists of two major phases: knowledge tuning and skills tuning. There are two steps in knowledge tuning where the first step learns simple knowledge (short samples) and the second step learns complicated knowledge (longer samples). The second step uses replay a replay buffer with data from the first step. Both foundational skills and compositional skills are learned during the skills tuning phases, where a replay buffer of data from the knowledge phase is used. Importantly, we use a set of hyper-parameters for training that are very different from standard small-scale supervised fine-training: larger batch size and carefully optimized learning rate and scheduler. ![Untitled](model-card/Model%20Card%20for%20Merlinite%207b%2028cc0b72cf574a4a828140d3539ede4a_Untitled%202.png) ## Model description - Language(s):* Primarily English - License: Apache 2.0 - Base model: [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) - Teacher Model: [mistralai/Mixtral-8x7B-Instruct-v0.1](https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1) ## Prompt Template ```python sys_prompt = "You are an AI language model developed by IBM Research. You are a cautious assistant. You carefully follow instructions. You are helpful and harmless and you follow ethical guidelines and promote positive behavior." prompt = f'<\|system\|>\n{sys_prompt}\n<\|user\|>\n{inputs}\n<\|assistant\|>\n' stop_token = '<\|endoftext\|>' ``` We advise utilizing the system prompt employed during the model's training for optimal inference performance, as there could be performance variations based on the provided instructions. ## Bias, Risks, and Limitations Merlinite-7b has not been aligned to human preferences, so the model might produce problematic outputs. The model might also maintain the limitations and constraints that arise from the base model. The model undergoes training on synthetic data, leading to the potential inheritance of both advantages and limitations from the underlying teacher models and data generation methods. The incorporation of safety measures during Merlinite-7b's training process is considered beneficial. However, a nuanced understanding of the associated risks requires detailed studies for more accurate quantification. In the absence of adequate safeguards and RLHF, there exists a risk of malicious utilization of these models for generating disinformation or harmful content. Caution is urged against complete reliance on a specific language model for crucial decisions or impactful information, as preventing these models from fabricating content is not straightforward. Additionally, it remains uncertain whether smaller models might exhibit increased susceptibility to hallucination in ungrounded generation scenarios due to their reduced sizes and memorization capacities. This aspect is currently an active area of research, and we anticipate more rigorous exploration, comprehension, and mitigations in this domain.	{"language": ["en"], "license": "apache-2.0", "tags": ["merlinite", "mistral", "ibm", "lab", "labrador", "labradorite"], "pipeline_tag": "text-generation", "base_model": "mistralai/Mistral-7B-v0.1"}	instructlab/merlinite-7b-lab	null	[ "transformers", "safetensors", "mistral", "text-generation", "merlinite", "ibm", "lab", "labrador", "labradorite", "conversational", "en", "arxiv:2403.01081", "base_model:mistralai/Mistral-7B-v0.1", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T03:00:10+00:00	[ "2403.01081" ]	[ "en" ]	TAGS #transformers #safetensors #mistral #text-generation #merlinite #ibm #lab #labrador #labradorite #conversational #en #arxiv-2403.01081 #base_model-mistralai/Mistral-7B-v0.1 #license-apache-2.0 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	Model Card for Merlinite 7b Paper ================================= ### Overview !Screenshot 2024-02-22 at 11.26.13 URL ### Performance [\] Numbers for models other than Merlinite-7b and Labradorite-13b (ours) are taken from lmsys/chatbot-arena-leaderboard [] Numbers taken from MistralAI Release Blog ### Method LAB: Large-scale Alignment for chatBots is a novel synthetic data-based alignment tuning method for LLMs from IBM Research. Merlinite-7b is a Mistral-7b-derivative model trained with the LAB methodology, using Mixtral-8x7b-Instruct as a teacher model. LAB consists of three key components: 1. Taxonomy-driven data curation process 2. Large-scale synthetic data generator 3. Two-phased-training with replay buffers !Untitled LAB approach allows for adding new knowledge and skills, in an incremental fashion, to an already pre-trained model without suffering from catastrophic forgetting. Taxonomy is a tree of seed examples that are used to prompt a teacher model to generate synthetic data. Taxonomy allows the data curator or the model designer to easily specify a diverse set of the knowledge-domains and skills that they would like to include in their LLM. At a high level, these can be categorized into three high-level bins - knowledge, foundational skills, and compositional skills. The leaf nodes of the taxonomy are tasks associated with one or more seed examples. !Untitled During the synthetic data generation, unlike previous approaches where seed examples are uniformly drawn from the entire pool (i.e. self-instruct), we use the taxonomy to drive the sampling process: For each knowledge/skill, we only use the local examples within the leaf node as seeds to prompt the teacher model. This makes the teacher model better exploit the task distributions defined by the local examples of each node and the diversity in the taxonomy itself ensures the entire generation covers a wide range of tasks, as illustrated below. In turns, this allows for using Mixtral 8x7B as the teacher model for generation while performing very competitively with models such as ORCA-2, WizardLM, and Zephyr Beta that rely on synthetic data generated by much larger and capable models like GPT-4. !URL For adding new domain-specific knowledge, we provide an external knowledge source (document) and prompt the model to generate questions and answers based on the document. Foundational skills such as reasoning and compositional skills such as creative writing are generated through in-context learning using the seed examples from the taxonomy. Additionally, to ensure the data is high-quality and safe, we employ steps to check the questions and answers to ensure that they are grounded and safe. This is done using the same teacher model that generated the data. Our training consists of two major phases: knowledge tuning and skills tuning. There are two steps in knowledge tuning where the first step learns simple knowledge (short samples) and the second step learns complicated knowledge (longer samples). The second step uses replay a replay buffer with data from the first step. Both foundational skills and compositional skills are learned during the skills tuning phases, where a replay buffer of data from the knowledge phase is used. Importantly, we use a set of hyper-parameters for training that are very different from standard small-scale supervised fine-training: larger batch size and carefully optimized learning rate and scheduler. !Untitled Model description ----------------- Language(s): Primarily English * License: Apache 2.0 * Base model: mistralai/Mistral-7B-v0.1 * Teacher Model: mistralai/Mixtral-8x7B-Instruct-v0.1 Prompt Template --------------- We advise utilizing the system prompt employed during the model's training for optimal inference performance, as there could be performance variations based on the provided instructions. Bias, Risks, and Limitations ---------------------------- Merlinite-7b has not been aligned to human preferences, so the model might produce problematic outputs. The model might also maintain the limitations and constraints that arise from the base model. The model undergoes training on synthetic data, leading to the potential inheritance of both advantages and limitations from the underlying teacher models and data generation methods. The incorporation of safety measures during Merlinite-7b's training process is considered beneficial. However, a nuanced understanding of the associated risks requires detailed studies for more accurate quantification. In the absence of adequate safeguards and RLHF, there exists a risk of malicious utilization of these models for generating disinformation or harmful content. Caution is urged against complete reliance on a specific language model for crucial decisions or impactful information, as preventing these models from fabricating content is not straightforward. Additionally, it remains uncertain whether smaller models might exhibit increased susceptibility to hallucination in ungrounded generation scenarios due to their reduced sizes and memorization capacities. This aspect is currently an active area of research, and we anticipate more rigorous exploration, comprehension, and mitigations in this domain.	[ "### Overview\n\n\n!Screenshot 2024-02-22 at 11.26.13 URL", "### Performance\n\n\n\n[\\] Numbers for models other than Merlinite-7b and Labradorite-13b (ours) are taken from lmsys/chatbot-arena-leaderboard\n\n\n[] Numbers taken from MistralAI Release Blog", "### Method\n\n\nLAB: Large-scale Alignment for chatBots is a novel synthetic data-based alignment tuning method for LLMs from IBM Research. Merlinite-7b is a Mistral-7b-derivative model trained with the LAB methodology, using Mixtral-8x7b-Instruct as a teacher model.\n\n\nLAB consists of three key components:\n\n\n1. Taxonomy-driven data curation process\n2. Large-scale synthetic data generator\n3. Two-phased-training with replay buffers\n\n\n!Untitled\n\n\nLAB approach allows for adding new knowledge and skills, in an incremental fashion, to an already pre-trained model without suffering from catastrophic forgetting.\n\n\nTaxonomy is a tree of seed examples that are used to prompt a teacher model to generate synthetic data. Taxonomy allows the data curator or the model designer to easily specify a diverse set of the knowledge-domains and skills that they would like to include in their LLM. At a high level, these can be categorized into three high-level bins - knowledge, foundational skills, and compositional skills. The leaf nodes of the taxonomy are tasks associated with one or more seed examples.\n\n\n!Untitled\n\n\nDuring the synthetic data generation, unlike previous approaches where seed examples are uniformly drawn from the entire pool (i.e. self-instruct), we use the taxonomy to drive the sampling process: For each knowledge/skill, we only use the local examples within the leaf node as seeds to prompt the teacher model.\nThis makes the teacher model better exploit the task distributions defined by the local examples of each node and the diversity in the taxonomy itself ensures the entire generation covers a wide range of tasks, as illustrated below. In turns, this allows for using Mixtral 8x7B as the teacher model for generation while performing very competitively with models such as ORCA-2, WizardLM, and Zephyr Beta that rely on synthetic data generated by much larger and capable models like GPT-4.\n\n\n!URL\n\n\nFor adding new domain-specific knowledge, we provide an external knowledge source (document) and prompt the model to generate questions and answers based on the document.\nFoundational skills such as reasoning and compositional skills such as creative writing are generated through in-context learning using the seed examples from the taxonomy.\n\n\nAdditionally, to ensure the data is high-quality and safe, we employ steps to check the questions and answers to ensure that they are grounded and safe. This is done using the same teacher model that generated the data.\n\n\nOur training consists of two major phases: knowledge tuning and skills tuning.\nThere are two steps in knowledge tuning where the first step learns simple knowledge (short samples) and the second step learns complicated knowledge (longer samples).\nThe second step uses replay a replay buffer with data from the first step.\nBoth foundational skills and compositional skills are learned during the skills tuning phases, where a replay buffer of data from the knowledge phase is used.\nImportantly, we use a set of hyper-parameters for training that are very different from standard small-scale supervised fine-training: larger batch size and carefully optimized learning rate and scheduler.\n\n\n!Untitled\n\n\nModel description\n-----------------\n\n\n Language(s): Primarily English\n* License: Apache 2.0\n* Base model: mistralai/Mistral-7B-v0.1\n* Teacher Model: mistralai/Mixtral-8x7B-Instruct-v0.1\n\n\nPrompt Template\n---------------\n\n\nWe advise utilizing the system prompt employed during the model's training for optimal inference performance, as there could be performance variations based on the provided instructions.\n\n\nBias, Risks, and Limitations\n----------------------------\n\n\nMerlinite-7b has not been aligned to human preferences, so the model might produce problematic outputs. The model might also maintain the limitations and constraints that arise from the base model.\n\n\nThe model undergoes training on synthetic data, leading to the potential inheritance of both advantages and limitations from the underlying teacher models and data generation methods. The incorporation of safety measures during Merlinite-7b's training process is considered beneficial. However, a nuanced understanding of the associated risks requires detailed studies for more accurate quantification.\n\n\nIn the absence of adequate safeguards and RLHF, there exists a risk of malicious utilization of these models for generating disinformation or harmful content. Caution is urged against complete reliance on a specific language model for crucial decisions or impactful information, as preventing these models from fabricating content is not straightforward. Additionally, it remains uncertain whether smaller models might exhibit increased susceptibility to hallucination in ungrounded generation scenarios due to their reduced sizes and memorization capacities. This aspect is currently an active area of research, and we anticipate more rigorous exploration, comprehension, and mitigations in this domain." ]	[ "TAGS\n#transformers #safetensors #mistral #text-generation #merlinite #ibm #lab #labrador #labradorite #conversational #en #arxiv-2403.01081 #base_model-mistralai/Mistral-7B-v0.1 #license-apache-2.0 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "### Overview\n\n\n!Screenshot 2024-02-22 at 11.26.13 URL", "### Performance\n\n\n\n[\\] Numbers for models other than Merlinite-7b and Labradorite-13b (ours) are taken from lmsys/chatbot-arena-leaderboard\n\n\n[] Numbers taken from MistralAI Release Blog", "### Method\n\n\nLAB: Large-scale Alignment for chatBots is a novel synthetic data-based alignment tuning method for LLMs from IBM Research. Merlinite-7b is a Mistral-7b-derivative model trained with the LAB methodology, using Mixtral-8x7b-Instruct as a teacher model.\n\n\nLAB consists of three key components:\n\n\n1. Taxonomy-driven data curation process\n2. Large-scale synthetic data generator\n3. Two-phased-training with replay buffers\n\n\n!Untitled\n\n\nLAB approach allows for adding new knowledge and skills, in an incremental fashion, to an already pre-trained model without suffering from catastrophic forgetting.\n\n\nTaxonomy is a tree of seed examples that are used to prompt a teacher model to generate synthetic data. Taxonomy allows the data curator or the model designer to easily specify a diverse set of the knowledge-domains and skills that they would like to include in their LLM. At a high level, these can be categorized into three high-level bins - knowledge, foundational skills, and compositional skills. The leaf nodes of the taxonomy are tasks associated with one or more seed examples.\n\n\n!Untitled\n\n\nDuring the synthetic data generation, unlike previous approaches where seed examples are uniformly drawn from the entire pool (i.e. self-instruct), we use the taxonomy to drive the sampling process: For each knowledge/skill, we only use the local examples within the leaf node as seeds to prompt the teacher model.\nThis makes the teacher model better exploit the task distributions defined by the local examples of each node and the diversity in the taxonomy itself ensures the entire generation covers a wide range of tasks, as illustrated below. In turns, this allows for using Mixtral 8x7B as the teacher model for generation while performing very competitively with models such as ORCA-2, WizardLM, and Zephyr Beta that rely on synthetic data generated by much larger and capable models like GPT-4.\n\n\n!URL\n\n\nFor adding new domain-specific knowledge, we provide an external knowledge source (document) and prompt the model to generate questions and answers based on the document.\nFoundational skills such as reasoning and compositional skills such as creative writing are generated through in-context learning using the seed examples from the taxonomy.\n\n\nAdditionally, to ensure the data is high-quality and safe, we employ steps to check the questions and answers to ensure that they are grounded and safe. This is done using the same teacher model that generated the data.\n\n\nOur training consists of two major phases: knowledge tuning and skills tuning.\nThere are two steps in knowledge tuning where the first step learns simple knowledge (short samples) and the second step learns complicated knowledge (longer samples).\nThe second step uses replay a replay buffer with data from the first step.\nBoth foundational skills and compositional skills are learned during the skills tuning phases, where a replay buffer of data from the knowledge phase is used.\nImportantly, we use a set of hyper-parameters for training that are very different from standard small-scale supervised fine-training: larger batch size and carefully optimized learning rate and scheduler.\n\n\n!Untitled\n\n\nModel description\n-----------------\n\n\n Language(s): Primarily English\n* License: Apache 2.0\n* Base model: mistralai/Mistral-7B-v0.1\n* Teacher Model: mistralai/Mixtral-8x7B-Instruct-v0.1\n\n\nPrompt Template\n---------------\n\n\nWe advise utilizing the system prompt employed during the model's training for optimal inference performance, as there could be performance variations based on the provided instructions.\n\n\nBias, Risks, and Limitations\n----------------------------\n\n\nMerlinite-7b has not been aligned to human preferences, so the model might produce problematic outputs. The model might also maintain the limitations and constraints that arise from the base model.\n\n\nThe model undergoes training on synthetic data, leading to the potential inheritance of both advantages and limitations from the underlying teacher models and data generation methods. The incorporation of safety measures during Merlinite-7b's training process is considered beneficial. However, a nuanced understanding of the associated risks requires detailed studies for more accurate quantification.\n\n\nIn the absence of adequate safeguards and RLHF, there exists a risk of malicious utilization of these models for generating disinformation or harmful content. Caution is urged against complete reliance on a specific language model for crucial decisions or impactful information, as preventing these models from fabricating content is not straightforward. Additionally, it remains uncertain whether smaller models might exhibit increased susceptibility to hallucination in ungrounded generation scenarios due to their reduced sizes and memorization capacities. This aspect is currently an active area of research, and we anticipate more rigorous exploration, comprehension, and mitigations in this domain." ]
null	transformers	# Merlinite 7b - GGUF 4-bit quantized version of [instructlab/merlinite-7b-lab](https://huggingface.co/instructlab/merlinite-7b-lab)	{"language": ["en"], "license": "apache-2.0", "tags": ["merlinite", "mistral", "ibm", "lab", "labrador", "labradorite"], "base_model": "instructlab/merlinite-7b-lab", "quantized_by": "IBM Research"}	instructlab/merlinite-7b-lab-GGUF	null	[ "transformers", "gguf", "mistral", "merlinite", "ibm", "lab", "labrador", "labradorite", "en", "base_model:instructlab/merlinite-7b-lab", "license:apache-2.0", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T03:00:27+00:00	[]	[ "en" ]	TAGS #transformers #gguf #mistral #merlinite #ibm #lab #labrador #labradorite #en #base_model-instructlab/merlinite-7b-lab #license-apache-2.0 #endpoints_compatible #text-generation-inference #region-us	# Merlinite 7b - GGUF 4-bit quantized version of instructlab/merlinite-7b-lab	[ "# Merlinite 7b - GGUF\n\n4-bit quantized version of instructlab/merlinite-7b-lab" ]	[ "TAGS\n#transformers #gguf #mistral #merlinite #ibm #lab #labrador #labradorite #en #base_model-instructlab/merlinite-7b-lab #license-apache-2.0 #endpoints_compatible #text-generation-inference #region-us \n", "# Merlinite 7b - GGUF\n\n4-bit quantized version of instructlab/merlinite-7b-lab" ]
text-generation	transformers	# Model Card for Granite-7b-lab [Paper](https://arxiv.org/abs/2403.01081) ### Overview ![Screenshot 2024-02-22 at 11.26.13 AM.png](model-card/Model%20Card%20for%20Merlinite%207b%2028cc0b72cf574a4a828140d3539ede4a_Screenshot_2024-02-22_at_11.26.13_AM.png) ### Performance \| Model \| Alignment \| Base \| Teacher \| MTBench (Avg) * \| MMLU(5-shot) \| \| --- \| --- \| --- \| --- \| --- \| --- \| \| [Llama-2-13b-chat-hf](https://huggingface.co/meta-llama/Llama-2-13b-chat-hf) \| RLHF \| Llama-2-13b \| Human Annotators \| 6.65 \|54.58 \| \| [Orca-2-13b](https://huggingface.co/microsoft/Orca-2-13b) \| Progressive Training \| Llama-2-13b \| GPT-4 \| 6.15 \| 60.37 * \| \| [WizardLM-13B-V1.2](https://huggingface.co/WizardLM/WizardLM-13B-V1.2) \| Evol-Instruct \| Llama-2-13b \| GPT-4 \| 7.20 \| 54.83 \| \| [Labradorite-13b](https://huggingface.co/ibm/labradorite-13b) \| Large-scale Alignment for chatBots (LAB) \| Llama-2-13b \| Mixtral-8x7B-Instruct \| 7.23 \| 58.89 \| \| [Mistral-7B-Instruct-v0.1](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.1) \| SFT \| Mistral-7B-v0.1 \| - \| 6.84 \| 60.37 \| \| [zephyr-7b-beta](https://huggingface.co/HuggingFaceH4/zephyr-7b-beta) \| SFT/DPO \| Mistral-7B-v0.1 \| GPT-4 \| 7.34 \| 61.07 \| \| [Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2) \| SFT \| Mistral-7B-v0.1 \| - \| 7.6** \| 60.78 \| \| [Merlinite-7b-lab](https://huggingface.co/instructlab/merlinite-7b-lab) \| Large-scale Alignment for chatBots (LAB) \| Mistral-7B-v0.1 \| Mixtral-8x7B-Instruct \| 7.66 \|64.88 \| \| Granite-7b-lab \| Large-scale Alignment for chatBots (LAB) \| Granite-7b-base\| Mixtral-8x7B-Instruct \| 6.69 \| 51.91 \| [] Numbers for models other than Merlinite-7b-lab, Granite-7b-lab and [Labradorite-13b](https://huggingface.co/ibm/labradorite-13b) are taken from [lmsys/chatbot-arena-leaderboard](https://huggingface.co/spaces/lmsys/chatbot-arena-leaderboard) [] Numbers taken from [MistralAI Release Blog](https://mistral.ai/news/la-plateforme/) ### Method LAB: Large-scale Alignment for chatBots is a novel synthetic data-based alignment tuning method for LLMs from IBM Research. Granite-7b-lab is a Granite-7b-base derivative model trained with the LAB methodology, using Mixtral-8x7b-Instruct as a teacher model. LAB consists of three key components: 1. Taxonomy-driven data curation process 2. Large-scale synthetic data generator 3. Two-phased-training with replay buffers ![Untitled](model-card/Model%20Card%20for%20Merlinite%207b%2028cc0b72cf574a4a828140d3539ede4a_Untitled.png) LAB approach allows for adding new knowledge and skills, in an incremental fashion, to an already pre-trained model without suffering from catastrophic forgetting. Taxonomy is a tree of seed examples that are used to prompt a teacher model to generate synthetic data. Taxonomy allows the data curator or the model designer to easily specify a diverse set of the knowledge-domains and skills that they would like to include in their LLM. At a high level, these can be categorized into three high-level bins - knowledge, foundational skills, and compositional skills. The leaf nodes of the taxonomy are tasks associated with one or more seed examples. ![Untitled](model-card/Model%20Card%20for%20Merlinite%207b%2028cc0b72cf574a4a828140d3539ede4a_Untitled%201.png) During the synthetic data generation, unlike previous approaches where seed examples are uniformly drawn from the entire pool (i.e. self-instruct), we use the taxonomy to drive the sampling process: For each knowledge/skill, we only use the local examples within the leaf node as seeds to prompt the teacher model. This makes the teacher model better exploit the task distributions defined by the local examples of each node and the diversity in the taxonomy itself ensures the entire generation covers a wide range of tasks, as illustrated below. In turns, this allows for using Mixtral 8x7B as the teacher model for generation while performing very competitively with models such as ORCA-2, WizardLM, and Zephyr Beta that rely on synthetic data generated by much larger and capable models like GPT-4. ![intuition.png](model-card/Model%20Card%20for%20Merlinite%207b%2028cc0b72cf574a4a828140d3539ede4a_intuition.png) For adding new domain-specific knowledge, we provide an external knowledge source (document) and prompt the model to generate questions and answers based on the document. Foundational skills such as reasoning and compositional skills such as creative writing are generated through in-context learning using the seed examples from the taxonomy. Additionally, to ensure the data is high-quality and safe, we employ steps to check the questions and answers to ensure that they are grounded and safe. This is done using the same teacher model that generated the data. Our training consists of two major phases: knowledge tuning and skills tuning. There are two steps in knowledge tuning where the first step learns simple knowledge (short samples) and the second step learns complicated knowledge (longer samples). The second step uses replay a replay buffer with data from the first step. Both foundational skills and compositional skills are learned during the skills tuning phases, where a replay buffer of data from the knowledge phase is used. Importantly, we use a set of hyper-parameters for training that are very different from standard small-scale supervised fine-training: larger batch size and carefully optimized learning rate and scheduler. ![Untitled](model-card/Model%20Card%20for%20Merlinite%207b%2028cc0b72cf574a4a828140d3539ede4a_Untitled%202.png) ## Model description - Model Name: Granite-7b-lab - Language(s):* Primarily English - License: Apache 2.0 - Base model: [ibm/granite-7b-base](https://huggingface.co/ibm/granite-7b-base) - Teacher Model: [mistralai/Mixtral-8x7B-Instruct-v0.1](https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1) ## Prompt Template ```python sys_prompt = "You are an AI language model developed by IBM Research. You are a cautious assistant. You carefully follow instructions. You are helpful and harmless and you follow ethical guidelines and promote positive behavior." prompt = f'<\|system\|>\n{sys_prompt}\n<\|user\|>\n{inputs}\n<\|assistant\|>\n' stop_token = '<\|endoftext\|>' ``` We advise utilizing the system prompt employed during the model's training for optimal inference performance, as there could be performance variations based on the provided instructions. Bias, Risks, and Limitations Granite-7b-lab is a base model and has not undergone any safety alignment, there it may produce problematic outputs. In the absence of adequate safeguards and RLHF, there exists a risk of malicious utilization of these models for generating disinformation or harmful content. Caution is urged against complete reliance on a specific language model for crucial decisions or impactful information, as preventing these models from fabricating content is not straightforward. Additionally, it remains uncertain whether smaller models might exhibit increased susceptibility to hallucination in ungrounded generation scenarios due to their reduced sizes and memorization capacities. This aspect is currently an active area of research, and we anticipate more rigorous exploration, comprehension, and mitigations in this domain.	{"language": ["en"], "license": "apache-2.0", "tags": ["granite", "ibm", "lab", "labrador", "labradorite"], "pipeline_tag": "text-generation", "base_model": "ibm/granite-7b-base"}	instructlab/granite-7b-lab	null	[ "transformers", "safetensors", "llama", "text-generation", "granite", "ibm", "lab", "labrador", "labradorite", "conversational", "en", "arxiv:2403.01081", "base_model:ibm/granite-7b-base", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T03:00:44+00:00	[ "2403.01081" ]	[ "en" ]	TAGS #transformers #safetensors #llama #text-generation #granite #ibm #lab #labrador #labradorite #conversational #en #arxiv-2403.01081 #base_model-ibm/granite-7b-base #license-apache-2.0 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	Model Card for Granite-7b-lab Paper =================================== ### Overview !Screenshot 2024-02-22 at 11.26.13 URL ### Performance [\] Numbers for models other than Merlinite-7b-lab, Granite-7b-lab and Labradorite-13b are taken from lmsys/chatbot-arena-leaderboard [] Numbers taken from MistralAI Release Blog ### Method LAB: Large-scale Alignment for chatBots is a novel synthetic data-based alignment tuning method for LLMs from IBM Research. Granite-7b-lab is a Granite-7b-base derivative model trained with the LAB methodology, using Mixtral-8x7b-Instruct as a teacher model. LAB consists of three key components: 1. Taxonomy-driven data curation process 2. Large-scale synthetic data generator 3. Two-phased-training with replay buffers !Untitled LAB approach allows for adding new knowledge and skills, in an incremental fashion, to an already pre-trained model without suffering from catastrophic forgetting. Taxonomy is a tree of seed examples that are used to prompt a teacher model to generate synthetic data. Taxonomy allows the data curator or the model designer to easily specify a diverse set of the knowledge-domains and skills that they would like to include in their LLM. At a high level, these can be categorized into three high-level bins - knowledge, foundational skills, and compositional skills. The leaf nodes of the taxonomy are tasks associated with one or more seed examples. !Untitled During the synthetic data generation, unlike previous approaches where seed examples are uniformly drawn from the entire pool (i.e. self-instruct), we use the taxonomy to drive the sampling process: For each knowledge/skill, we only use the local examples within the leaf node as seeds to prompt the teacher model. This makes the teacher model better exploit the task distributions defined by the local examples of each node and the diversity in the taxonomy itself ensures the entire generation covers a wide range of tasks, as illustrated below. In turns, this allows for using Mixtral 8x7B as the teacher model for generation while performing very competitively with models such as ORCA-2, WizardLM, and Zephyr Beta that rely on synthetic data generated by much larger and capable models like GPT-4. !URL For adding new domain-specific knowledge, we provide an external knowledge source (document) and prompt the model to generate questions and answers based on the document. Foundational skills such as reasoning and compositional skills such as creative writing are generated through in-context learning using the seed examples from the taxonomy. Additionally, to ensure the data is high-quality and safe, we employ steps to check the questions and answers to ensure that they are grounded and safe. This is done using the same teacher model that generated the data. Our training consists of two major phases: knowledge tuning and skills tuning. There are two steps in knowledge tuning where the first step learns simple knowledge (short samples) and the second step learns complicated knowledge (longer samples). The second step uses replay a replay buffer with data from the first step. Both foundational skills and compositional skills are learned during the skills tuning phases, where a replay buffer of data from the knowledge phase is used. Importantly, we use a set of hyper-parameters for training that are very different from standard small-scale supervised fine-training: larger batch size and carefully optimized learning rate and scheduler. !Untitled Model description ----------------- Model Name: Granite-7b-lab * Language(s): Primarily English * License: Apache 2.0 * Base model: ibm/granite-7b-base * Teacher Model: mistralai/Mixtral-8x7B-Instruct-v0.1 Prompt Template --------------- We advise utilizing the system prompt employed during the model's training for optimal inference performance, as there could be performance variations based on the provided instructions. Bias, Risks, and Limitations Granite-7b-lab is a base model and has not undergone any safety alignment, there it may produce problematic outputs. In the absence of adequate safeguards and RLHF, there exists a risk of malicious utilization of these models for generating disinformation or harmful content. Caution is urged against complete reliance on a specific language model for crucial decisions or impactful information, as preventing these models from fabricating content is not straightforward. Additionally, it remains uncertain whether smaller models might exhibit increased susceptibility to hallucination in ungrounded generation scenarios due to their reduced sizes and memorization capacities. This aspect is currently an active area of research, and we anticipate more rigorous exploration, comprehension, and mitigations in this domain.	[ "### Overview\n\n\n!Screenshot 2024-02-22 at 11.26.13 URL", "### Performance\n\n\n\n[\\] Numbers for models other than Merlinite-7b-lab, Granite-7b-lab and Labradorite-13b are taken from lmsys/chatbot-arena-leaderboard\n\n\n[] Numbers taken from MistralAI Release Blog", "### Method\n\n\nLAB: Large-scale Alignment for chatBots is a novel synthetic data-based alignment tuning method for LLMs from IBM Research. Granite-7b-lab is a Granite-7b-base derivative model trained with the LAB methodology, using Mixtral-8x7b-Instruct as a teacher model.\n\n\nLAB consists of three key components:\n\n\n1. Taxonomy-driven data curation process\n2. Large-scale synthetic data generator\n3. Two-phased-training with replay buffers\n\n\n!Untitled\n\n\nLAB approach allows for adding new knowledge and skills, in an incremental fashion, to an already pre-trained model without suffering from catastrophic forgetting.\n\n\nTaxonomy is a tree of seed examples that are used to prompt a teacher model to generate synthetic data. Taxonomy allows the data curator or the model designer to easily specify a diverse set of the knowledge-domains and skills that they would like to include in their LLM. At a high level, these can be categorized into three high-level bins - knowledge, foundational skills, and compositional skills. The leaf nodes of the taxonomy are tasks associated with one or more seed examples.\n\n\n!Untitled\n\n\nDuring the synthetic data generation, unlike previous approaches where seed examples are uniformly drawn from the entire pool (i.e. self-instruct), we use the taxonomy to drive the sampling process: For each knowledge/skill, we only use the local examples within the leaf node as seeds to prompt the teacher model.\nThis makes the teacher model better exploit the task distributions defined by the local examples of each node and the diversity in the taxonomy itself ensures the entire generation covers a wide range of tasks, as illustrated below. In turns, this allows for using Mixtral 8x7B as the teacher model for generation while performing very competitively with models such as ORCA-2, WizardLM, and Zephyr Beta that rely on synthetic data generated by much larger and capable models like GPT-4.\n\n\n!URL\n\n\nFor adding new domain-specific knowledge, we provide an external knowledge source (document) and prompt the model to generate questions and answers based on the document.\nFoundational skills such as reasoning and compositional skills such as creative writing are generated through in-context learning using the seed examples from the taxonomy.\n\n\nAdditionally, to ensure the data is high-quality and safe, we employ steps to check the questions and answers to ensure that they are grounded and safe. This is done using the same teacher model that generated the data.\n\n\nOur training consists of two major phases: knowledge tuning and skills tuning.\nThere are two steps in knowledge tuning where the first step learns simple knowledge (short samples) and the second step learns complicated knowledge (longer samples).\nThe second step uses replay a replay buffer with data from the first step.\nBoth foundational skills and compositional skills are learned during the skills tuning phases, where a replay buffer of data from the knowledge phase is used.\nImportantly, we use a set of hyper-parameters for training that are very different from standard small-scale supervised fine-training: larger batch size and carefully optimized learning rate and scheduler.\n\n\n!Untitled\n\n\nModel description\n-----------------\n\n\n Model Name: Granite-7b-lab\n* Language(s): Primarily English\n* License: Apache 2.0\n* Base model: ibm/granite-7b-base\n* Teacher Model: mistralai/Mixtral-8x7B-Instruct-v0.1\n\n\nPrompt Template\n---------------\n\n\nWe advise utilizing the system prompt employed during the model's training for optimal inference performance, as there could be performance variations based on the provided instructions.\n\n\nBias, Risks, and Limitations\n\n\nGranite-7b-lab is a base model and has not undergone any safety alignment, there it may produce problematic outputs. In the absence of adequate safeguards and RLHF, there exists a risk of malicious utilization of these models for generating disinformation or harmful content. Caution is urged against complete reliance on a specific language model for crucial decisions or impactful information, as preventing these models from fabricating content is not straightforward. Additionally, it remains uncertain whether smaller models might exhibit increased susceptibility to hallucination in ungrounded generation scenarios due to their reduced sizes and memorization capacities. This aspect is currently an active area of research, and we anticipate more rigorous exploration, comprehension, and mitigations in this domain." ]	[ "TAGS\n#transformers #safetensors #llama #text-generation #granite #ibm #lab #labrador #labradorite #conversational #en #arxiv-2403.01081 #base_model-ibm/granite-7b-base #license-apache-2.0 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "### Overview\n\n\n!Screenshot 2024-02-22 at 11.26.13 URL", "### Performance\n\n\n\n[\\] Numbers for models other than Merlinite-7b-lab, Granite-7b-lab and Labradorite-13b are taken from lmsys/chatbot-arena-leaderboard\n\n\n[] Numbers taken from MistralAI Release Blog", "### Method\n\n\nLAB: Large-scale Alignment for chatBots is a novel synthetic data-based alignment tuning method for LLMs from IBM Research. Granite-7b-lab is a Granite-7b-base derivative model trained with the LAB methodology, using Mixtral-8x7b-Instruct as a teacher model.\n\n\nLAB consists of three key components:\n\n\n1. Taxonomy-driven data curation process\n2. Large-scale synthetic data generator\n3. Two-phased-training with replay buffers\n\n\n!Untitled\n\n\nLAB approach allows for adding new knowledge and skills, in an incremental fashion, to an already pre-trained model without suffering from catastrophic forgetting.\n\n\nTaxonomy is a tree of seed examples that are used to prompt a teacher model to generate synthetic data. Taxonomy allows the data curator or the model designer to easily specify a diverse set of the knowledge-domains and skills that they would like to include in their LLM. At a high level, these can be categorized into three high-level bins - knowledge, foundational skills, and compositional skills. The leaf nodes of the taxonomy are tasks associated with one or more seed examples.\n\n\n!Untitled\n\n\nDuring the synthetic data generation, unlike previous approaches where seed examples are uniformly drawn from the entire pool (i.e. self-instruct), we use the taxonomy to drive the sampling process: For each knowledge/skill, we only use the local examples within the leaf node as seeds to prompt the teacher model.\nThis makes the teacher model better exploit the task distributions defined by the local examples of each node and the diversity in the taxonomy itself ensures the entire generation covers a wide range of tasks, as illustrated below. In turns, this allows for using Mixtral 8x7B as the teacher model for generation while performing very competitively with models such as ORCA-2, WizardLM, and Zephyr Beta that rely on synthetic data generated by much larger and capable models like GPT-4.\n\n\n!URL\n\n\nFor adding new domain-specific knowledge, we provide an external knowledge source (document) and prompt the model to generate questions and answers based on the document.\nFoundational skills such as reasoning and compositional skills such as creative writing are generated through in-context learning using the seed examples from the taxonomy.\n\n\nAdditionally, to ensure the data is high-quality and safe, we employ steps to check the questions and answers to ensure that they are grounded and safe. This is done using the same teacher model that generated the data.\n\n\nOur training consists of two major phases: knowledge tuning and skills tuning.\nThere are two steps in knowledge tuning where the first step learns simple knowledge (short samples) and the second step learns complicated knowledge (longer samples).\nThe second step uses replay a replay buffer with data from the first step.\nBoth foundational skills and compositional skills are learned during the skills tuning phases, where a replay buffer of data from the knowledge phase is used.\nImportantly, we use a set of hyper-parameters for training that are very different from standard small-scale supervised fine-training: larger batch size and carefully optimized learning rate and scheduler.\n\n\n!Untitled\n\n\nModel description\n-----------------\n\n\n Model Name: Granite-7b-lab\n* Language(s): Primarily English\n* License: Apache 2.0\n* Base model: ibm/granite-7b-base\n* Teacher Model: mistralai/Mixtral-8x7B-Instruct-v0.1\n\n\nPrompt Template\n---------------\n\n\nWe advise utilizing the system prompt employed during the model's training for optimal inference performance, as there could be performance variations based on the provided instructions.\n\n\nBias, Risks, and Limitations\n\n\nGranite-7b-lab is a base model and has not undergone any safety alignment, there it may produce problematic outputs. In the absence of adequate safeguards and RLHF, there exists a risk of malicious utilization of these models for generating disinformation or harmful content. Caution is urged against complete reliance on a specific language model for crucial decisions or impactful information, as preventing these models from fabricating content is not straightforward. Additionally, it remains uncertain whether smaller models might exhibit increased susceptibility to hallucination in ungrounded generation scenarios due to their reduced sizes and memorization capacities. This aspect is currently an active area of research, and we anticipate more rigorous exploration, comprehension, and mitigations in this domain." ]
text-generation	peft	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # results This model is a fine-tuned version of [NousResearch/Llama-2-7b-chat-hf](https://huggingface.co/NousResearch/Llama-2-7b-chat-hf) on the generator dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 4 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: constant - lr_scheduler_warmup_ratio: 0.03 - num_epochs: 1 ### Framework versions - PEFT 0.10.0 - Transformers 4.38.2 - Pytorch 2.2.1+cu121 - Datasets 2.18.0 - Tokenizers 0.15.2	{"library_name": "peft", "tags": ["trl", "sft", "generated_from_trainer"], "datasets": ["generator"], "base_model": "NousResearch/Llama-2-7b-chat-hf", "pipeline_tag": "text-generation", "model-index": [{"name": "results", "results": []}]}	AWeirdDev/zh-tw-llama-2	null	[ "peft", "safetensors", "trl", "sft", "generated_from_trainer", "text-generation", "dataset:generator", "base_model:NousResearch/Llama-2-7b-chat-hf", "region:us" ]	null	2024-04-19T03:00:50+00:00	[]	[]	TAGS #peft #safetensors #trl #sft #generated_from_trainer #text-generation #dataset-generator #base_model-NousResearch/Llama-2-7b-chat-hf #region-us	# results This model is a fine-tuned version of NousResearch/Llama-2-7b-chat-hf on the generator dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 4 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: constant - lr_scheduler_warmup_ratio: 0.03 - num_epochs: 1 ### Framework versions - PEFT 0.10.0 - Transformers 4.38.2 - Pytorch 2.2.1+cu121 - Datasets 2.18.0 - Tokenizers 0.15.2	[ "# results\n\nThis model is a fine-tuned version of NousResearch/Llama-2-7b-chat-hf on the generator dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 0.0002\n- train_batch_size: 4\n- eval_batch_size: 8\n- seed: 42\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: constant\n- lr_scheduler_warmup_ratio: 0.03\n- num_epochs: 1", "### Framework versions\n\n- PEFT 0.10.0\n- Transformers 4.38.2\n- Pytorch 2.2.1+cu121\n- Datasets 2.18.0\n- Tokenizers 0.15.2" ]	[ "TAGS\n#peft #safetensors #trl #sft #generated_from_trainer #text-generation #dataset-generator #base_model-NousResearch/Llama-2-7b-chat-hf #region-us \n", "# results\n\nThis model is a fine-tuned version of NousResearch/Llama-2-7b-chat-hf on the generator dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 0.0002\n- train_batch_size: 4\n- eval_batch_size: 8\n- seed: 42\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: constant\n- lr_scheduler_warmup_ratio: 0.03\n- num_epochs: 1", "### Framework versions\n\n- PEFT 0.10.0\n- Transformers 4.38.2\n- Pytorch 2.2.1+cu121\n- Datasets 2.18.0\n- Tokenizers 0.15.2" ]
null	transformers	# Granite 7b - GGUF 4-bit quantized version of [instructlab/granite-7b-lab](https://huggingface.co/instructlab/granite-7b-lab)	{"language": ["en"], "license": "apache-2.0", "tags": ["granite", "ibm", "lab", "labrador", "labradorite"], "base_model": "instructlab/granite-7b-lab", "quantized_by": "IBM Research"}	instructlab/granite-7b-lab-GGUF	null	[ "transformers", "gguf", "llama", "granite", "ibm", "lab", "labrador", "labradorite", "en", "base_model:instructlab/granite-7b-lab", "license:apache-2.0", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T03:00:54+00:00	[]	[ "en" ]	TAGS #transformers #gguf #llama #granite #ibm #lab #labrador #labradorite #en #base_model-instructlab/granite-7b-lab #license-apache-2.0 #endpoints_compatible #text-generation-inference #region-us	# Granite 7b - GGUF 4-bit quantized version of instructlab/granite-7b-lab	[ "# Granite 7b - GGUF\n\n4-bit quantized version of instructlab/granite-7b-lab" ]	[ "TAGS\n#transformers #gguf #llama #granite #ibm #lab #labrador #labradorite #en #base_model-instructlab/granite-7b-lab #license-apache-2.0 #endpoints_compatible #text-generation-inference #region-us \n", "# Granite 7b - GGUF\n\n4-bit quantized version of instructlab/granite-7b-lab" ]
text-generation	transformers	# Clevyby/flammen19X-mistral-7B-Q5_K_S-GGUF This model was converted to GGUF format from [`flammenai/flammen19X-mistral-7B`](https://huggingface.co/flammenai/flammen19X-mistral-7B) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space. Refer to the [original model card](https://huggingface.co/flammenai/flammen19X-mistral-7B) for more details on the model. ### Note: The additional files in this GGUF repo is for personal usage using Text Gen Webui with llamacpp_hf.	{"library_name": "transformers", "tags": ["llama-cpp", "gguf-my-repo"]}	Clevyby/flammen19X-mistral-7B-Q5_K_S-GGUF	null	[ "transformers", "gguf", "mistral", "text-generation", "llama-cpp", "gguf-my-repo", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T03:00:55+00:00	[]	[]	TAGS #transformers #gguf #mistral #text-generation #llama-cpp #gguf-my-repo #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	# Clevyby/flammen19X-mistral-7B-Q5_K_S-GGUF This model was converted to GGUF format from 'flammenai/flammen19X-mistral-7B' using URL via the URL's GGUF-my-repo space. Refer to the original model card for more details on the model. ### Note: The additional files in this GGUF repo is for personal usage using Text Gen Webui with llamacpp_hf.	[ "# Clevyby/flammen19X-mistral-7B-Q5_K_S-GGUF\nThis model was converted to GGUF format from 'flammenai/flammen19X-mistral-7B' using URL via the URL's GGUF-my-repo space.\nRefer to the original model card for more details on the model.", "### Note: \nThe additional files in this GGUF repo is for personal usage using Text Gen Webui with llamacpp_hf." ]	[ "TAGS\n#transformers #gguf #mistral #text-generation #llama-cpp #gguf-my-repo #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "# Clevyby/flammen19X-mistral-7B-Q5_K_S-GGUF\nThis model was converted to GGUF format from 'flammenai/flammen19X-mistral-7B' using URL via the URL's GGUF-my-repo space.\nRefer to the original model card for more details on the model.", "### Note: \nThe additional files in this GGUF repo is for personal usage using Text Gen Webui with llamacpp_hf." ]
fill-mask	transformers	# 5CD-AI/visobert-14gb-corpus ## Overview <!-- Provide a quick summary of what the model is/does. --> We continually pretrain `uitnlp/visobert` on a merged 14GB dataset, the training dataset includes: - Internal data (100M comments and 15M posts on Facebook) - UIT data, which is used to pretrain `uitnlp/visobert` - MC4 ecommerce Here are the results on 4 downstream tasks on Vietnamese social media texts, including Emotion Recognition(UIT-VSMEC), Hate Speech Detection(UIT-HSD), Spam Reviews Detection(ViSpamReviews), Hate Speech Spans Detection(ViHOS): <table> <tr align="center"> <td rowspan=2><b>Model</td> <td rowspan=2><b>Avg</td> <td colspan=3><b>Emotion Recognition</td> <td colspan=3><b>Hate Speech Detection</td> <td colspan=3><b>Spam Reviews Detection</td> <td colspan=3><b>Hate Speech Spans Detection</td> </tr> <tr align="center"> <td><b>Acc</td> <td><b>WF1</td> <td><b>MF1</td> <td><b>Acc</td> <td><b>WF1</td> <td><b>MF1</td> <td><b>Acc</td> <td><b>WF1</td> <td><b>MF1</td> <td><b>Acc</td> <td><b>WF1</td> <td><b>MF1</td> </tr> <tr align="center"> <td align="left">viBERT</td> <td>78.16</td> <td>61.91</td> <td>61.98</td> <td>59.7</td> <td>85.34</td> <td>85.01</td> <td>62.07</td> <td>89.93</td> <td>89.79</td> <td>76.8</td> <td>90.42</td> <td>90.45</td> <td>84.55</td> </tr> <tr align="center"> <td align="left">vELECTRA</td> <td>79.23</td> <td>64.79</td> <td>64.71</td> <td>61.95</td> <td>86.96</td> <td>86.37</td> <td>63.95</td> <td>89.83</td> <td>89.68</td> <td>76.23</td> <td>90.59</td> <td>90.58</td> <td>85.12</td> </tr> <tr align="center"> <td align="left">PhoBERT-Base </td> <td>79.3</td> <td>63.49</td> <td>63.36</td> <td>61.41</td> <td>87.12</td> <td>86.81</td> <td>65.01</td> <td>89.83</td> <td>89.75</td> <td>76.18</td> <td>91.32</td> <td>91.38</td> <td>85.92</td> </tr> <tr align="center"> <td align="left">PhoBERT-Large</td> <td>79.82</td> <td>64.71</td> <td>64.66</td> <td>62.55</td> <td>87.32</td> <td>86.98</td> <td>65.14</td> <td>90.12</td> <td>90.03</td> <td>76.88</td> <td>91.44</td> <td>91.46</td> <td>86.56</td> </tr> <tr align="center"> <td align="left">ViSoBERT</td> <td>81.58</td> <td>68.1</td> <td>68.37</td> <td>65.88</td> <td>88.51</td> <td>88.31</td> <td>68.77</td> <td>90.99</td> <td><b>90.92</td> <td><b>79.06</td> <td>91.62</td> <td>91.57</td> <td>86.8</td> </tr> <tr align="center"> <td align="left">visobert-14gb-corpus</td> <td><b>82.2</td> <td><b>68.69</td> <td><b>68.75</td> <td><b>66.03</td> <td><b>88.79</td> <td><b>88.6</td> <td><b>69.57</td> <td><b>91.02</td> <td>90.88</td> <td>77.13</td> <td><b>93.69</td> <td><b>93.63</td> <td><b>89.66</td> </tr> </div> </table> ## Usage (HuggingFace Transformers) Install `transformers` package: pip install transformers Then you can use this model for fill-mask task like this: ```python from transformers import pipeline model_path = "5CD-AI/visobert-14gb-corpus" mask_filler = pipeline("fill-mask", model_path) mask_filler("ăn nói xà <mask>", top_k=10) ``` ## Fine-tune Configuration We fine-tune `5CD-AI/visobert-14gb-corpus` on 4 downstream tasks with `transformers` library with the following configuration: - seed: 42 - gradient_accumulation_steps: 1 - weight_decay: 0.01 - optimizer: AdamW with betas=(0.9, 0.999) and epsilon=1e-08 - training_epochs: 30 - model_max_length: 128 - learning_rate: 1e-5 - metric_for_best_model: wf1 - strategy: epoch And different additional configurations for each task: \| Emotion Recognition \| Hate Speech Detection \| Spam Reviews Detection \| Hate Speech Spans Detection \| \| --------------------------------------------------------------------------------- \| --------------------------------------------------------------------------------- \| --------------------------------------------------------------------------------- \| --------------------------------------------------------------------------------- \| \|\- train_batch_size: 64<br>\- lr_scheduler_type: linear \| \- train_batch_size: 32<br>\- lr_scheduler_type: linear \| \- train_batch_size: 32<br>\- lr_scheduler_type: cosine \| \- train_batch_size: 32<br>\- lr_scheduler_type: cosine \|	{"library_name": "transformers", "tags": [], "pipeline_tag": "fill-mask", "widget": [{"text": "shop l\u00e0m \u0103n nh\u01b0 c\u00e1i <mask>"}, {"text": "hag t\u1eeb Qu\u1ea3ng <mask> k\u1ef1c n\u00e9t"}, {"text": "Set xinh qu\u00e1, <mask> b\u00e8o nh\u00e8o"}, {"text": "\u0103n n\u00f3i x\u00e0 <mask>"}]}	5CD-AI/visobert-14gb-corpus	null	[ "transformers", "safetensors", "xlm-roberta", "fill-mask", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T03:01:37+00:00	[]	[]	TAGS #transformers #safetensors #xlm-roberta #fill-mask #autotrain_compatible #endpoints_compatible #region-us	5CD-AI/visobert-14gb-corpus =========================== Overview -------- We continually pretrain 'uitnlp/visobert' on a merged 14GB dataset, the training dataset includes: * Internal data (100M comments and 15M posts on Facebook) * UIT data, which is used to pretrain 'uitnlp/visobert' * MC4 ecommerce Here are the results on 4 downstream tasks on Vietnamese social media texts, including Emotion Recognition(UIT-VSMEC), Hate Speech Detection(UIT-HSD), Spam Reviews Detection(ViSpamReviews), Hate Speech Spans Detection(ViHOS): Usage (HuggingFace Transformers) -------------------------------- Install 'transformers' package: ``` pip install transformers ``` Then you can use this model for fill-mask task like this: Fine-tune Configuration ----------------------- We fine-tune '5CD-AI/visobert-14gb-corpus' on 4 downstream tasks with 'transformers' library with the following configuration: * seed: 42 * gradient\_accumulation\_steps: 1 * weight\_decay: 0.01 * optimizer: AdamW with betas=(0.9, 0.999) and epsilon=1e-08 * training\_epochs: 30 * model\_max\_length: 128 * learning\_rate: 1e-5 * metric\_for\_best\_model: wf1 * strategy: epoch And different additional configurations for each task:	[]	[ "TAGS\n#transformers #safetensors #xlm-roberta #fill-mask #autotrain_compatible #endpoints_compatible #region-us \n" ]
text-generation	transformers	There currently is an issue with the model generating random reserved special tokens (like "<\|reserved_special_token_49\|>") at the end. Please use with `skip_special_tokens=true`. We will update once we found the reason for this behaviour. If you found a solution, please let us know! # Llama 3 DiscoLM German 8b v0.1 Experimental <p align="center"><img src="disco_llama.webp" width="400"></p> # Introduction Llama 3 DiscoLM German 8b v0.1 Experimental is an experimental Llama 3 based version of [DiscoLM German](https://huggingface.co/DiscoResearch/DiscoLM_German_7b_v1). This is an experimental release and not intended for production use. The model is still in development and will be updated with new features and improvements in the future. Please find a online Demo [here](https://364b61f772fa7baacb.gradio.live/) (we may take this offline for updates). # Prompt Format DiscoLM German uses ChatML as the prompt format which enables OpenAI endpoint compatability and is supported by most inference libraries and frontends. System prompts allow steerability and interesting new ways to interact with an LLM, guiding rules, roles, and stylistic choices of the model. ``` <\|im_start\|>system Du bist ein hilfreicher Assistent.<\|im_end\|> <\|im_start\|>user Wer bist du?<\|im_end\|> <\|im_start\|>assistant Ich bin ein Sprachmodell namens DiscoLM German und ich wurde von DiscoResearch trainiert.<\|im_end\|> ``` This prompt is available as a [chat template](https://huggingface.co/docs/transformers/main/chat_templating), which means you can format messages using the `tokenizer.apply_chat_template()` method: ```python messages = [ {"role": "system", "content": "Du bist ein hilfreicher Assistent."}, {"role": "user", "content": "Wer bist du?"} ] gen_input = tokenizer.apply_chat_template(message, return_tensors="pt") model.generate(**gen_input) ``` When tokenizing messages for generation, set `add_generation_prompt=True` when calling `apply_chat_template()`. This will append `<\|im_start\|>assistant\n` to your prompt, to ensure that the model continues with an assistant response. # Example Code for Inference ```python model_id = "DiscoResearch/Llama3_DiscoLM_German_8b_v0.1_experimental" tokenizer = AutoTokenizer.from_pretrained(model_id) model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.bfloat16, device_map="auto", ) messages = [ {"role": "system", "content": "Du bist ein hilfreicher Assistent."}, {"role": "user", "content": "Wer bist du?"}, ] input_ids = tokenizer.apply_chat_template( messages, add_generation_prompt=True, return_tensors="pt" ).to(model.device) terminators = [ tokenizer.eos_token_id, tokenizer.convert_tokens_to_ids("<\|eot_id\|>") ] outputs = model.generate( input_ids, max_new_tokens=256, eos_token_id=terminators, do_sample=True, temperature=0.6, top_p=0.9, ) response = outputs[0][input_ids.shape[-1]:] print(tokenizer.decode(response, skip_special_tokens=True)) ``` # Limitations & Biases This model can produce factually incorrect and offensive output, and should not be relied on to produce factually accurate information. This model was trained on various public datasets. While great efforts have been taken to clean the pretraining data, it is possible that this model could generate biased or otherwise offensive outputs and it is the responsibility of the user to implement a safety/moderation layer. Please use with caution. # License This model is distributed under the META LLAMA 3 COMMUNITY LICENSE, see [LICENSE](LICENSE) for more information. # Acknowledgements Built with Meta Llama 3. DiscoLM German is a [DiscoResearch](https://huggingface.co/DiscoResearch) project, a collective effort by [JP Harries](https://huggingface.co/jphme), [Björn Plüster](https://huggingface.co/bjoernp) and [Daniel Auras](https://huggingface.co/rasdani). Development of Llama 3 DiscoLM German 8b was sponsored by [ellamind](https://ellamind.com). Compute was sponsored generously by [sysGen GmbH](https://www.sysgen.de/). [<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl) # About DiscoResearch DiscoResearch is an aspiring open research community for AI enthusiasts and LLM hackers. Come join our [Discord](https://discord.gg/ttNdas89f3), share your opinions and ideas, and advance open LLM research with us! # Disclaimer The license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. This model should only be deployed with additional safety measures in place.	{"library_name": "transformers", "tags": []}	DiscoResearch/Llama3_DiscoLM_German_8b_v0.1_experimental	null	[ "transformers", "safetensors", "llama", "text-generation", "conversational", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T03:02:32+00:00	[]	[]	TAGS #transformers #safetensors #llama #text-generation #conversational #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	There currently is an issue with the model generating random reserved special tokens (like "<\|reserved_special_token_49\|>") at the end. Please use with 'skip_special_tokens=true'. We will update once we found the reason for this behaviour. If you found a solution, please let us know! # Llama 3 DiscoLM German 8b v0.1 Experimental <p align="center"><img src="disco_llama.webp" width="400"></p> # Introduction Llama 3 DiscoLM German 8b v0.1 Experimental is an experimental Llama 3 based version of DiscoLM German. This is an experimental release and not intended for production use. The model is still in development and will be updated with new features and improvements in the future. Please find a online Demo here (we may take this offline for updates). # Prompt Format DiscoLM German uses ChatML as the prompt format which enables OpenAI endpoint compatability and is supported by most inference libraries and frontends. System prompts allow steerability and interesting new ways to interact with an LLM, guiding rules, roles, and stylistic choices of the model. This prompt is available as a chat template, which means you can format messages using the 'tokenizer.apply_chat_template()' method: When tokenizing messages for generation, set 'add_generation_prompt=True' when calling 'apply_chat_template()'. This will append '<\|im_start\|>assistant\n' to your prompt, to ensure that the model continues with an assistant response. # Example Code for Inference # Limitations & Biases This model can produce factually incorrect and offensive output, and should not be relied on to produce factually accurate information. This model was trained on various public datasets. While great efforts have been taken to clean the pretraining data, it is possible that this model could generate biased or otherwise offensive outputs and it is the responsibility of the user to implement a safety/moderation layer. Please use with caution. # License This model is distributed under the META LLAMA 3 COMMUNITY LICENSE, see LICENSE for more information. # Acknowledgements Built with Meta Llama 3. DiscoLM German is a DiscoResearch project, a collective effort by JP Harries, Björn Plüster and Daniel Auras. Development of Llama 3 DiscoLM German 8b was sponsored by ellamind. Compute was sponsored generously by sysGen GmbH. <img src="URL alt="Built with Axolotl" width="200" height="32"/> # About DiscoResearch DiscoResearch is an aspiring open research community for AI enthusiasts and LLM hackers. Come join our Discord, share your opinions and ideas, and advance open LLM research with us! # Disclaimer The license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. This model should only be deployed with additional safety measures in place.	[ "# Llama 3 DiscoLM German 8b v0.1 Experimental\n\n<p align=\"center\"><img src=\"disco_llama.webp\" width=\"400\"></p>", "# Introduction\n\nLlama 3 DiscoLM German 8b v0.1 Experimental is an experimental Llama 3 based version of DiscoLM German.\n\nThis is an experimental release and not intended for production use. The model is still in development and will be updated with new features and improvements in the future.\n\nPlease find a online Demo here (we may take this offline for updates).", "# Prompt Format\n\nDiscoLM German uses ChatML as the prompt format which enables OpenAI endpoint compatability and is supported by most inference libraries and frontends.\n\nSystem prompts allow steerability and interesting new ways to interact with an LLM, guiding rules, roles, and stylistic choices of the model.\n\n\n\nThis prompt is available as a chat template, which means you can format messages using the\n'tokenizer.apply_chat_template()' method:\n\n\n\nWhen tokenizing messages for generation, set 'add_generation_prompt=True' when calling 'apply_chat_template()'. This will append '<\|im_start\|>assistant\\n' to your prompt, to ensure\nthat the model continues with an assistant response.", "# Example Code for Inference", "# Limitations & Biases\n\nThis model can produce factually incorrect and offensive output, and should not be relied on to produce factually accurate information.\nThis model was trained on various public datasets. While great efforts have been taken to clean the pretraining data, it is possible that this model could generate biased or otherwise offensive outputs and it is the responsibility of the user to implement a safety/moderation layer. Please use with caution.", "# License\n\nThis model is distributed under the META LLAMA 3 COMMUNITY LICENSE, see LICENSE for more information.", "# Acknowledgements\n\nBuilt with Meta Llama 3.\n\nDiscoLM German is a DiscoResearch project, a collective effort by JP Harries, Björn Plüster and Daniel Auras.\n\nDevelopment of Llama 3 DiscoLM German 8b was sponsored by ellamind.\nCompute was sponsored generously by sysGen GmbH.\n\n<img src=\"URL alt=\"Built with Axolotl\" width=\"200\" height=\"32\"/>", "# About DiscoResearch\n\nDiscoResearch is an aspiring open research community for AI enthusiasts and LLM hackers. Come join our Discord, share your opinions and ideas, and advance open LLM research with us!", "# Disclaimer\n\nThe license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. This model should only be deployed with additional safety measures in place." ]	[ "TAGS\n#transformers #safetensors #llama #text-generation #conversational #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "# Llama 3 DiscoLM German 8b v0.1 Experimental\n\n<p align=\"center\"><img src=\"disco_llama.webp\" width=\"400\"></p>", "# Introduction\n\nLlama 3 DiscoLM German 8b v0.1 Experimental is an experimental Llama 3 based version of DiscoLM German.\n\nThis is an experimental release and not intended for production use. The model is still in development and will be updated with new features and improvements in the future.\n\nPlease find a online Demo here (we may take this offline for updates).", "# Prompt Format\n\nDiscoLM German uses ChatML as the prompt format which enables OpenAI endpoint compatability and is supported by most inference libraries and frontends.\n\nSystem prompts allow steerability and interesting new ways to interact with an LLM, guiding rules, roles, and stylistic choices of the model.\n\n\n\nThis prompt is available as a chat template, which means you can format messages using the\n'tokenizer.apply_chat_template()' method:\n\n\n\nWhen tokenizing messages for generation, set 'add_generation_prompt=True' when calling 'apply_chat_template()'. This will append '<\|im_start\|>assistant\\n' to your prompt, to ensure\nthat the model continues with an assistant response.", "# Example Code for Inference", "# Limitations & Biases\n\nThis model can produce factually incorrect and offensive output, and should not be relied on to produce factually accurate information.\nThis model was trained on various public datasets. While great efforts have been taken to clean the pretraining data, it is possible that this model could generate biased or otherwise offensive outputs and it is the responsibility of the user to implement a safety/moderation layer. Please use with caution.", "# License\n\nThis model is distributed under the META LLAMA 3 COMMUNITY LICENSE, see LICENSE for more information.", "# Acknowledgements\n\nBuilt with Meta Llama 3.\n\nDiscoLM German is a DiscoResearch project, a collective effort by JP Harries, Björn Plüster and Daniel Auras.\n\nDevelopment of Llama 3 DiscoLM German 8b was sponsored by ellamind.\nCompute was sponsored generously by sysGen GmbH.\n\n<img src=\"URL alt=\"Built with Axolotl\" width=\"200\" height=\"32\"/>", "# About DiscoResearch\n\nDiscoResearch is an aspiring open research community for AI enthusiasts and LLM hackers. Come join our Discord, share your opinions and ideas, and advance open LLM research with us!", "# Disclaimer\n\nThe license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. This model should only be deployed with additional safety measures in place." ]
text-generation	transformers	# Model Trained Using AutoTrain This model was trained using AutoTrain. For more information, please visit [AutoTrain](https://hf.co/docs/autotrain). # Usage ```python from transformers import AutoModelForCausalLM, AutoTokenizer model_path = "PATH_TO_THIS_REPO" tokenizer = AutoTokenizer.from_pretrained(model_path) model = AutoModelForCausalLM.from_pretrained( model_path, device_map="auto", torch_dtype='auto' ).eval() # Prompt content: "hi" messages = [ {"role": "user", "content": "hi"} ] input_ids = tokenizer.apply_chat_template(conversation=messages, tokenize=True, add_generation_prompt=True, return_tensors='pt') output_ids = model.generate(input_ids.to('cuda')) response = tokenizer.decode(output_ids[0][input_ids.shape[1]:], skip_special_tokens=True) # Model response: "Hello! How can I assist you today?" print(response) ```	{"license": "other", "library_name": "transformers", "tags": ["autotrain", "text-generation-inference", "text-generation", "peft"], "widget": [{"messages": [{"role": "user", "content": "What is your favorite condiment?"}]}]}	archangel268/autotrain-kbwyx-5jxtc	null	[ "transformers", "tensorboard", "safetensors", "autotrain", "text-generation-inference", "text-generation", "peft", "conversational", "license:other", "endpoints_compatible", "has_space", "region:us" ]	null	2024-04-19T03:03:39+00:00	[]	[]	TAGS #transformers #tensorboard #safetensors #autotrain #text-generation-inference #text-generation #peft #conversational #license-other #endpoints_compatible #has_space #region-us	# Model Trained Using AutoTrain This model was trained using AutoTrain. For more information, please visit AutoTrain. # Usage	[ "# Model Trained Using AutoTrain\n\nThis model was trained using AutoTrain. For more information, please visit AutoTrain.", "# Usage" ]	[ "TAGS\n#transformers #tensorboard #safetensors #autotrain #text-generation-inference #text-generation #peft #conversational #license-other #endpoints_compatible #has_space #region-us \n", "# Model Trained Using AutoTrain\n\nThis model was trained using AutoTrain. For more information, please visit AutoTrain.", "# Usage" ]
text-generation	transformers	# NikolayKozloff/Mistral-portuguese-luana-7b-Mathematics-Q8_0-GGUF This model was converted to GGUF format from [`rhaymison/Mistral-portuguese-luana-7b-Mathematics`](https://huggingface.co/rhaymison/Mistral-portuguese-luana-7b-Mathematics) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space. Refer to the [original model card](https://huggingface.co/rhaymison/Mistral-portuguese-luana-7b-Mathematics) for more details on the model. ## Use with llama.cpp Install llama.cpp through brew. ```bash brew install ggerganov/ggerganov/llama.cpp ``` Invoke the llama.cpp server or the CLI. CLI: ```bash llama-cli --hf-repo NikolayKozloff/Mistral-portuguese-luana-7b-Mathematics-Q8_0-GGUF --model mistral-portuguese-luana-7b-mathematics.Q8_0.gguf -p "The meaning to life and the universe is" ``` Server: ```bash llama-server --hf-repo NikolayKozloff/Mistral-portuguese-luana-7b-Mathematics-Q8_0-GGUF --model mistral-portuguese-luana-7b-mathematics.Q8_0.gguf -c 2048 ``` Note: You can also use this checkpoint directly through the [usage steps](https://github.com/ggerganov/llama.cpp?tab=readme-ov-file#usage) listed in the Llama.cpp repo as well. ``` git clone https://github.com/ggerganov/llama.cpp && cd llama.cpp && make && ./main -m mistral-portuguese-luana-7b-mathematics.Q8_0.gguf -n 128 ```	{"language": ["pt"], "license": "apache-2.0", "library_name": "transformers", "tags": ["portuguese", "math", "mathematics", "matematica", "llama-cpp", "gguf-my-repo"], "datasets": ["rhaymison/orca-math-portuguese-64k"], "base_model": "rhaymison/Mistral-portuguese-luana-7b", "pipeline_tag": "text-generation", "model-index": [{"name": "Mistral-portuguese-luana-7b-Mathematics", "results": [{"task": {"type": "text-generation", "name": "Text Generation"}, "dataset": {"name": "ENEM Challenge (No Images)", "type": "eduagarcia/enem_challenge", "split": "train", "args": {"num_few_shot": 3}}, "metrics": [{"type": "acc", "value": 56.68, "name": "accuracy"}], "source": {"url": "https://huggingface.co/spaces/eduagarcia/open_pt_llm_leaderboard?query=rhaymison/Mistral-portuguese-luana-7b-Mathematics", "name": "Open Portuguese LLM Leaderboard"}}, {"task": {"type": "text-generation", "name": "Text Generation"}, "dataset": {"name": "BLUEX (No Images)", "type": "eduagarcia-temp/BLUEX_without_images", "split": "train", "args": {"num_few_shot": 3}}, "metrics": [{"type": "acc", "value": 45.9, "name": "accuracy"}], "source": {"url": "https://huggingface.co/spaces/eduagarcia/open_pt_llm_leaderboard?query=rhaymison/Mistral-portuguese-luana-7b-Mathematics", "name": "Open Portuguese LLM Leaderboard"}}, {"task": {"type": "text-generation", "name": "Text Generation"}, "dataset": {"name": "OAB Exams", "type": "eduagarcia/oab_exams", "split": "train", "args": {"num_few_shot": 3}}, "metrics": [{"type": "acc", "value": 37.9, "name": "accuracy"}], "source": {"url": "https://huggingface.co/spaces/eduagarcia/open_pt_llm_leaderboard?query=rhaymison/Mistral-portuguese-luana-7b-Mathematics", "name": "Open Portuguese LLM Leaderboard"}}, {"task": {"type": "text-generation", "name": "Text Generation"}, "dataset": {"name": "Assin2 RTE", "type": "assin2", "split": "test", "args": {"num_few_shot": 15}}, "metrics": [{"type": "f1_macro", "value": 89.36, "name": "f1-macro"}], "source": {"url": "https://huggingface.co/spaces/eduagarcia/open_pt_llm_leaderboard?query=rhaymison/Mistral-portuguese-luana-7b-Mathematics", "name": "Open Portuguese LLM Leaderboard"}}, {"task": {"type": "text-generation", "name": "Text Generation"}, "dataset": {"name": "Assin2 STS", "type": "eduagarcia/portuguese_benchmark", "split": "test", "args": {"num_few_shot": 15}}, "metrics": [{"type": "pearson", "value": 74.78, "name": "pearson"}], "source": {"url": "https://huggingface.co/spaces/eduagarcia/open_pt_llm_leaderboard?query=rhaymison/Mistral-portuguese-luana-7b-Mathematics", "name": "Open Portuguese LLM Leaderboard"}}, {"task": {"type": "text-generation", "name": "Text Generation"}, "dataset": {"name": "FaQuAD NLI", "type": "ruanchaves/faquad-nli", "split": "test", "args": {"num_few_shot": 15}}, "metrics": [{"type": "f1_macro", "value": 74.87, "name": "f1-macro"}], "source": {"url": "https://huggingface.co/spaces/eduagarcia/open_pt_llm_leaderboard?query=rhaymison/Mistral-portuguese-luana-7b-Mathematics", "name": "Open Portuguese LLM Leaderboard"}}, {"task": {"type": "text-generation", "name": "Text Generation"}, "dataset": {"name": "HateBR Binary", "type": "ruanchaves/hatebr", "split": "test", "args": {"num_few_shot": 25}}, "metrics": [{"type": "f1_macro", "value": 76.39, "name": "f1-macro"}], "source": {"url": "https://huggingface.co/spaces/eduagarcia/open_pt_llm_leaderboard?query=rhaymison/Mistral-portuguese-luana-7b-Mathematics", "name": "Open Portuguese LLM Leaderboard"}}, {"task": {"type": "text-generation", "name": "Text Generation"}, "dataset": {"name": "PT Hate Speech Binary", "type": "hate_speech_portuguese", "split": "test", "args": {"num_few_shot": 25}}, "metrics": [{"type": "f1_macro", "value": 67.46, "name": "f1-macro"}], "source": {"url": "https://huggingface.co/spaces/eduagarcia/open_pt_llm_leaderboard?query=rhaymison/Mistral-portuguese-luana-7b-Mathematics", "name": "Open Portuguese LLM Leaderboard"}}, {"task": {"type": "text-generation", "name": "Text Generation"}, "dataset": {"name": "tweetSentBR", "type": "eduagarcia/tweetsentbr_fewshot", "split": "test", "args": {"num_few_shot": 25}}, "metrics": [{"type": "f1_macro", "value": 49.03, "name": "f1-macro"}], "source": {"url": "https://huggingface.co/spaces/eduagarcia/open_pt_llm_leaderboard?query=rhaymison/Mistral-portuguese-luana-7b-Mathematics", "name": "Open Portuguese LLM Leaderboard"}}]}]}	NikolayKozloff/Mistral-portuguese-luana-7b-Mathematics-GGUF	null	[ "transformers", "gguf", "portuguese", "math", "mathematics", "matematica", "llama-cpp", "gguf-my-repo", "text-generation", "pt", "dataset:rhaymison/orca-math-portuguese-64k", "base_model:rhaymison/Mistral-portuguese-luana-7b", "license:apache-2.0", "model-index", "endpoints_compatible", "region:us" ]	null	2024-04-19T03:03:41+00:00	[]	[ "pt" ]	TAGS #transformers #gguf #portuguese #math #mathematics #matematica #llama-cpp #gguf-my-repo #text-generation #pt #dataset-rhaymison/orca-math-portuguese-64k #base_model-rhaymison/Mistral-portuguese-luana-7b #license-apache-2.0 #model-index #endpoints_compatible #region-us	# NikolayKozloff/Mistral-portuguese-luana-7b-Mathematics-Q8_0-GGUF This model was converted to GGUF format from 'rhaymison/Mistral-portuguese-luana-7b-Mathematics' using URL via the URL's GGUF-my-repo space. Refer to the original model card for more details on the model. ## Use with URL Install URL through brew. Invoke the URL server or the CLI. CLI: Server: Note: You can also use this checkpoint directly through the usage steps listed in the URL repo as well.	[ "# NikolayKozloff/Mistral-portuguese-luana-7b-Mathematics-Q8_0-GGUF\nThis model was converted to GGUF format from 'rhaymison/Mistral-portuguese-luana-7b-Mathematics' using URL via the URL's GGUF-my-repo space.\nRefer to the original model card for more details on the model.", "## Use with URL\n\nInstall URL through brew.\n\n\nInvoke the URL server or the CLI.\n\nCLI:\n\n\n\nServer:\n\n\n\nNote: You can also use this checkpoint directly through the usage steps listed in the URL repo as well." ]	[ "TAGS\n#transformers #gguf #portuguese #math #mathematics #matematica #llama-cpp #gguf-my-repo #text-generation #pt #dataset-rhaymison/orca-math-portuguese-64k #base_model-rhaymison/Mistral-portuguese-luana-7b #license-apache-2.0 #model-index #endpoints_compatible #region-us \n", "# NikolayKozloff/Mistral-portuguese-luana-7b-Mathematics-Q8_0-GGUF\nThis model was converted to GGUF format from 'rhaymison/Mistral-portuguese-luana-7b-Mathematics' using URL via the URL's GGUF-my-repo space.\nRefer to the original model card for more details on the model.", "## Use with URL\n\nInstall URL through brew.\n\n\nInvoke the URL server or the CLI.\n\nCLI:\n\n\n\nServer:\n\n\n\nNote: You can also use this checkpoint directly through the usage steps listed in the URL repo as well." ]
text-classification	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # electra-distilled-sst This model is a fine-tuned version of [google/electra-small-discriminator](https://huggingface.co/google/electra-small-discriminator) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 0.3772 - Accuracy: 0.9174 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 4.854751380955516e-05 - train_batch_size: 128 - eval_batch_size: 128 - seed: 33 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 9 - mixed_precision_training: Native AMP ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\| \| 0.3762 \| 1.0 \| 527 \| 0.2865 \| 0.9083 \| \| 0.2105 \| 2.0 \| 1054 \| 0.3472 \| 0.9048 \| \| 0.1566 \| 3.0 \| 1581 \| 0.3098 \| 0.9243 \| \| 0.1221 \| 4.0 \| 2108 \| 0.3681 \| 0.9174 \| \| 0.1029 \| 5.0 \| 2635 \| 0.3772 \| 0.9174 \| ### Framework versions - Transformers 4.40.0 - Pytorch 2.2.0 - Datasets 2.18.0 - Tokenizers 0.19.1	{"license": "apache-2.0", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "google/electra-small-discriminator", "model-index": [{"name": "electra-distilled-sst", "results": []}]}	kasohrab/electra-distilled-sst	null	[ "transformers", "tensorboard", "safetensors", "electra", "text-classification", "generated_from_trainer", "base_model:google/electra-small-discriminator", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T03:06:12+00:00	[]	[]	TAGS #transformers #tensorboard #safetensors #electra #text-classification #generated_from_trainer #base_model-google/electra-small-discriminator #license-apache-2.0 #autotrain_compatible #endpoints_compatible #region-us	electra-distilled-sst ===================== This model is a fine-tuned version of google/electra-small-discriminator on an unknown dataset. It achieves the following results on the evaluation set: * Loss: 0.3772 * Accuracy: 0.9174 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * learning\_rate: 4.854751380955516e-05 * train\_batch\_size: 128 * eval\_batch\_size: 128 * seed: 33 * optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 * lr\_scheduler\_type: linear * num\_epochs: 9 * mixed\_precision\_training: Native AMP ### Training results ### Framework versions * Transformers 4.40.0 * Pytorch 2.2.0 * Datasets 2.18.0 * Tokenizers 0.19.1	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 4.854751380955516e-05\n* train\\_batch\\_size: 128\n* eval\\_batch\\_size: 128\n* seed: 33\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* num\\_epochs: 9\n* mixed\\_precision\\_training: Native AMP", "### Training results", "### Framework versions\n\n\n* Transformers 4.40.0\n* Pytorch 2.2.0\n* Datasets 2.18.0\n* Tokenizers 0.19.1" ]	[ "TAGS\n#transformers #tensorboard #safetensors #electra #text-classification #generated_from_trainer #base_model-google/electra-small-discriminator #license-apache-2.0 #autotrain_compatible #endpoints_compatible #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 4.854751380955516e-05\n* train\\_batch\\_size: 128\n* eval\\_batch\\_size: 128\n* seed: 33\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* num\\_epochs: 9\n* mixed\\_precision\\_training: Native AMP", "### Training results", "### Framework versions\n\n\n* Transformers 4.40.0\n* Pytorch 2.2.0\n* Datasets 2.18.0\n* Tokenizers 0.19.1" ]
reinforcement-learning	stable-baselines3	# A2C Agent playing PandaReachDense-v3 This is a trained model of a A2C agent playing PandaReachDense-v3 using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3). ## Usage (with Stable-baselines3) TODO: Add your code ```python from stable_baselines3 import ... from huggingface_sb3 import load_from_hub ... ```	{"library_name": "stable-baselines3", "tags": ["PandaReachDense-v3", "deep-reinforcement-learning", "reinforcement-learning", "stable-baselines3"], "model-index": [{"name": "A2C", "results": [{"task": {"type": "reinforcement-learning", "name": "reinforcement-learning"}, "dataset": {"name": "PandaReachDense-v3", "type": "PandaReachDense-v3"}, "metrics": [{"type": "mean_reward", "value": "-0.22 +/- 0.11", "name": "mean_reward", "verified": false}]}]}]}	Rudolph314/a2c-PandaReachDense-v3	null	[ "stable-baselines3", "PandaReachDense-v3", "deep-reinforcement-learning", "reinforcement-learning", "model-index", "region:us" ]	null	2024-04-19T03:07:14+00:00	[]	[]	TAGS #stable-baselines3 #PandaReachDense-v3 #deep-reinforcement-learning #reinforcement-learning #model-index #region-us	# A2C Agent playing PandaReachDense-v3 This is a trained model of a A2C agent playing PandaReachDense-v3 using the stable-baselines3 library. ## Usage (with Stable-baselines3) TODO: Add your code	[ "# A2C Agent playing PandaReachDense-v3\nThis is a trained model of a A2C agent playing PandaReachDense-v3\nusing the stable-baselines3 library.", "## Usage (with Stable-baselines3)\nTODO: Add your code" ]	[ "TAGS\n#stable-baselines3 #PandaReachDense-v3 #deep-reinforcement-learning #reinforcement-learning #model-index #region-us \n", "# A2C Agent playing PandaReachDense-v3\nThis is a trained model of a A2C agent playing PandaReachDense-v3\nusing the stable-baselines3 library.", "## Usage (with Stable-baselines3)\nTODO: Add your code" ]
text-generation	transformers	# LYNN - AI for Roleplay <img src="./reallynn.png" alt="it's lynn!" width="340"/> > [!TIP] > This model is overfitted to the role-playing dataset; normal conversations may not work well. # Soliloquy-L3 Soliloquy-L3 is a fast, highly capable roleplaying model designed for immersive, dynamic experiences. Trained on over 250 million tokens of roleplaying data, Soliloquy-L3 has a vast knowledge base, rich literary expression, and support for up to 24k context length. It outperforms existing ~13B models, delivering enhanced roleplaying capabilities. ## Model Info \| Context Length \| Parameter \| Prompt Template \| isErp \| \| --- \| --- \| --- \| --- \| \| 24k(24576) \| 8B \| Llama 3 Chat \| Partly \| ## Prompt Template Use can you following jinja2 template. Which is identical to chat_template in [tokenizer_config](./tokenizer_config.json). ``` {% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<\|start_header_id\|>' + message['role'] + '<\|end_header_id\|>\n\n'+ message['content'] \| trim + '<\|eot_id\|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<\|start_header_id\|>assistant<\|end_header_id\|>\n\n' }}{% endif %} ``` ## License This model is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International Public License, under [META LLAMA 3 COMMUNITY LICENSE AGREEMENT](https://llama.meta.com/llama3/license/) If you would like to use this model for commercial purposes, please use our proprietary API. (Currently avilable at OpenRouter) For non-commercial use, please adhere to the terms of the CC BY-NC-SA 4.0 license. You are free to share and adapt the model for non-commercial purposes, provided you give appropriate credit, indicate if changes were made, and do not imply endorsement by the licensor. For more information about the CC BY-NC 4.0 license, please visit: https://creativecommons.org/licenses/by-nc-sa/4.0/ If you have any questions or would like to inquire about licensing, please contact us. ## Llama 3 Intended Use Intended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. Out-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English. Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. [https://llama.meta.com/llama3/license](https://llama.meta.com/llama3/license) ## Join our Discord [Join LYNN Discord](https://discord.gg/xuZVqUyG4Y)	{"language": ["en"], "license": "cc-by-nc-sa-4.0"}	openlynn/Llama-3-Soliloquy-8B-v1	null	[ "transformers", "safetensors", "llama", "text-generation", "conversational", "en", "license:cc-by-nc-sa-4.0", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T03:11:31+00:00	[]	[ "en" ]	TAGS #transformers #safetensors #llama #text-generation #conversational #en #license-cc-by-nc-sa-4.0 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	LYNN - AI for Roleplay ====================== ![it's lynn!](./URL) > > [!TIP] > This model is overfitted to the role-playing dataset; normal conversations may not work well. > > > Soliloquy-L3 ============ Soliloquy-L3 is a fast, highly capable roleplaying model designed for immersive, dynamic experiences. Trained on over 250 million tokens of roleplaying data, Soliloquy-L3 has a vast knowledge base, rich literary expression, and support for up to 24k context length. It outperforms existing ~13B models, delivering enhanced roleplaying capabilities. Model Info ---------- Prompt Template --------------- Use can you following jinja2 template. Which is identical to chat\_template in tokenizer\_config. License ------- This model is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International Public License, under META LLAMA 3 COMMUNITY LICENSE AGREEMENT If you would like to use this model for commercial purposes, please use our proprietary API. (Currently avilable at OpenRouter) For non-commercial use, please adhere to the terms of the CC BY-NC-SA 4.0 license. You are free to share and adapt the model for non-commercial purposes, provided you give appropriate credit, indicate if changes were made, and do not imply endorsement by the licensor. For more information about the CC BY-NC 4.0 license, please visit: URL If you have any questions or would like to inquire about licensing, please contact us. Llama 3 Intended Use -------------------- Intended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. Out-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English. Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. URL Join our Discord ---------------- Join LYNN Discord	[]	[ "TAGS\n#transformers #safetensors #llama #text-generation #conversational #en #license-cc-by-nc-sa-4.0 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n" ]
fill-mask	transformers	<!-- This model card has been generated automatically according to the information Keras had access to. You should probably proofread and complete it, then remove this comment. --> # jfforero/distilbert-base-uncased-BERT-POET2 This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset. It achieves the following results on the evaluation set: - Train Loss: 5.1534 - Validation Loss: 5.9706 - Epoch: 0 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'module': 'transformers.optimization_tf', 'class_name': 'WarmUp', 'config': {'initial_learning_rate': 2e-05, 'decay_schedule_fn': {'module': 'keras.optimizers.schedules', 'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': -998, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered_name': None}, 'warmup_steps': 1000, 'power': 1.0, 'name': None}, 'registered_name': 'WarmUp'}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.01} - training_precision: float32 ### Training results \| Train Loss \| Validation Loss \| Epoch \| \|:----------:\|:---------------:\|:-----:\| \| 5.1534 \| 5.9706 \| 0 \| ### Framework versions - Transformers 4.38.2 - TensorFlow 2.15.0 - Datasets 2.18.0 - Tokenizers 0.15.2	{"license": "apache-2.0", "tags": ["generated_from_keras_callback"], "base_model": "distilbert-base-uncased", "model-index": [{"name": "jfforero/distilbert-base-uncased-BERT-POET2", "results": []}]}	jfforero/distilbert-base-uncased-BERT-POET2	null	[ "transformers", "tf", "distilbert", "fill-mask", "generated_from_keras_callback", "base_model:distilbert-base-uncased", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T03:12:08+00:00	[]	[]	TAGS #transformers #tf #distilbert #fill-mask #generated_from_keras_callback #base_model-distilbert-base-uncased #license-apache-2.0 #autotrain_compatible #endpoints_compatible #region-us	jfforero/distilbert-base-uncased-BERT-POET2 =========================================== This model is a fine-tuned version of distilbert-base-uncased on an unknown dataset. It achieves the following results on the evaluation set: * Train Loss: 5.1534 * Validation Loss: 5.9706 * Epoch: 0 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * optimizer: {'name': 'AdamWeightDecay', 'learning\_rate': {'module': 'transformers.optimization\_tf', 'class\_name': 'WarmUp', 'config': {'initial\_learning\_rate': 2e-05, 'decay\_schedule\_fn': {'module': 'keras.optimizers.schedules', 'class\_name': 'PolynomialDecay', 'config': {'initial\_learning\_rate': 2e-05, 'decay\_steps': -998, 'end\_learning\_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered\_name': None}, 'warmup\_steps': 1000, 'power': 1.0, 'name': None}, 'registered\_name': 'WarmUp'}, 'decay': 0.0, 'beta\_1': 0.9, 'beta\_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight\_decay\_rate': 0.01} * training\_precision: float32 ### Training results ### Framework versions * Transformers 4.38.2 * TensorFlow 2.15.0 * Datasets 2.18.0 * Tokenizers 0.15.2	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* optimizer: {'name': 'AdamWeightDecay', 'learning\\_rate': {'module': 'transformers.optimization\\_tf', 'class\\_name': 'WarmUp', 'config': {'initial\\_learning\\_rate': 2e-05, 'decay\\_schedule\\_fn': {'module': 'keras.optimizers.schedules', 'class\\_name': 'PolynomialDecay', 'config': {'initial\\_learning\\_rate': 2e-05, 'decay\\_steps': -998, 'end\\_learning\\_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered\\_name': None}, 'warmup\\_steps': 1000, 'power': 1.0, 'name': None}, 'registered\\_name': 'WarmUp'}, 'decay': 0.0, 'beta\\_1': 0.9, 'beta\\_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight\\_decay\\_rate': 0.01}\n* training\\_precision: float32", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* TensorFlow 2.15.0\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #tf #distilbert #fill-mask #generated_from_keras_callback #base_model-distilbert-base-uncased #license-apache-2.0 #autotrain_compatible #endpoints_compatible #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* optimizer: {'name': 'AdamWeightDecay', 'learning\\_rate': {'module': 'transformers.optimization\\_tf', 'class\\_name': 'WarmUp', 'config': {'initial\\_learning\\_rate': 2e-05, 'decay\\_schedule\\_fn': {'module': 'keras.optimizers.schedules', 'class\\_name': 'PolynomialDecay', 'config': {'initial\\_learning\\_rate': 2e-05, 'decay\\_steps': -998, 'end\\_learning\\_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered\\_name': None}, 'warmup\\_steps': 1000, 'power': 1.0, 'name': None}, 'registered\\_name': 'WarmUp'}, 'decay': 0.0, 'beta\\_1': 0.9, 'beta\\_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight\\_decay\\_rate': 0.01}\n* training\\_precision: float32", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* TensorFlow 2.15.0\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]
text-generation	transformers	# Clevyby/Silicon-Alice-7B-Q5_K_S-GGUF This model was converted to GGUF format from [`LakoMoor/Silicon-Alice-7B`](https://huggingface.co/LakoMoor/Silicon-Alice-7B) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space. Refer to the [original model card](https://huggingface.co/LakoMoor/Silicon-Alice-7B) for more details on the model. ### Note: The additional files in this GGUF repo is for personal usage using Text Gen Webui with llamacpp_hf.	{"language": ["en", "ru"], "license": "cc-by-nc-4.0", "library_name": "transformers", "tags": ["not-for-all-audiences", "nsfw", "llama-cpp", "gguf-my-repo"], "model_name": "Silicon-Alice-7B", "base_model": ["LakoMoor/Silicon-Masha-7B"], "inference": false, "model_creator": "LakoMoor", "model_type": "mistral"}	Clevyby/Silicon-Alice-7B-Q5_K_S-GGUF	null	[ "transformers", "gguf", "mistral", "text-generation", "not-for-all-audiences", "nsfw", "llama-cpp", "gguf-my-repo", "en", "ru", "base_model:LakoMoor/Silicon-Masha-7B", "license:cc-by-nc-4.0", "autotrain_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T03:17:09+00:00	[]	[ "en", "ru" ]	TAGS #transformers #gguf #mistral #text-generation #not-for-all-audiences #nsfw #llama-cpp #gguf-my-repo #en #ru #base_model-LakoMoor/Silicon-Masha-7B #license-cc-by-nc-4.0 #autotrain_compatible #text-generation-inference #region-us	# Clevyby/Silicon-Alice-7B-Q5_K_S-GGUF This model was converted to GGUF format from 'LakoMoor/Silicon-Alice-7B' using URL via the URL's GGUF-my-repo space. Refer to the original model card for more details on the model. ### Note: The additional files in this GGUF repo is for personal usage using Text Gen Webui with llamacpp_hf.	[ "# Clevyby/Silicon-Alice-7B-Q5_K_S-GGUF\nThis model was converted to GGUF format from 'LakoMoor/Silicon-Alice-7B' using URL via the URL's GGUF-my-repo space.\nRefer to the original model card for more details on the model.", "### Note: \nThe additional files in this GGUF repo is for personal usage using Text Gen Webui with llamacpp_hf." ]	[ "TAGS\n#transformers #gguf #mistral #text-generation #not-for-all-audiences #nsfw #llama-cpp #gguf-my-repo #en #ru #base_model-LakoMoor/Silicon-Masha-7B #license-cc-by-nc-4.0 #autotrain_compatible #text-generation-inference #region-us \n", "# Clevyby/Silicon-Alice-7B-Q5_K_S-GGUF\nThis model was converted to GGUF format from 'LakoMoor/Silicon-Alice-7B' using URL via the URL's GGUF-my-repo space.\nRefer to the original model card for more details on the model.", "### Note: \nThe additional files in this GGUF repo is for personal usage using Text Gen Webui with llamacpp_hf." ]
fill-mask	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # tapt_helpfulness_base_pretraining_no_condencing This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 1.3379 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0001 - train_batch_size: 32 - eval_batch_size: 64 - seed: 42 - gradient_accumulation_steps: 32 - total_train_batch_size: 1024 - optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-06 - lr_scheduler_type: linear - num_epochs: 10 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| \|:-------------:\|:-----:\|:----:\|:---------------:\| \| 1.6371 \| 1.0 \| 112 \| 1.4525 \| \| 1.5107 \| 2.0 \| 225 \| 1.4102 \| \| 1.4871 \| 2.99 \| 337 \| 1.3845 \| \| 1.445 \| 4.0 \| 450 \| 1.3778 \| \| 1.4412 \| 4.99 \| 562 \| 1.3483 \| \| 1.4062 \| 6.0 \| 675 \| 1.3400 \| \| 1.4028 \| 6.99 \| 787 \| 1.3190 \| \| 1.3837 \| 8.0 \| 900 \| 1.3237 \| \| 1.3734 \| 9.0 \| 1013 \| 1.3322 \| \| 1.3725 \| 9.95 \| 1120 \| 1.3289 \| ### Framework versions - Transformers 4.30.2 - Pytorch 1.13.1+cu117 - Datasets 2.13.2 - Tokenizers 0.13.3	{"license": "mit", "tags": ["generated_from_trainer"], "model-index": [{"name": "tapt_helpfulness_base_pretraining_no_condencing", "results": []}]}	ltuzova/tapt_helpfulness_base_pretraining_no_condencing	null	[ "transformers", "pytorch", "roberta", "fill-mask", "generated_from_trainer", "license:mit", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T03:18:52+00:00	[]	[]	TAGS #transformers #pytorch #roberta #fill-mask #generated_from_trainer #license-mit #autotrain_compatible #endpoints_compatible #region-us	tapt\_helpfulness\_base\_pretraining\_no\_condencing ==================================================== This model is a fine-tuned version of roberta-base on an unknown dataset. It achieves the following results on the evaluation set: * Loss: 1.3379 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * learning\_rate: 0.0001 * train\_batch\_size: 32 * eval\_batch\_size: 64 * seed: 42 * gradient\_accumulation\_steps: 32 * total\_train\_batch\_size: 1024 * optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-06 * lr\_scheduler\_type: linear * num\_epochs: 10 ### Training results ### Framework versions * Transformers 4.30.2 * Pytorch 1.13.1+cu117 * Datasets 2.13.2 * Tokenizers 0.13.3	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 0.0001\n* train\\_batch\\_size: 32\n* eval\\_batch\\_size: 64\n* seed: 42\n* gradient\\_accumulation\\_steps: 32\n* total\\_train\\_batch\\_size: 1024\n* optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-06\n* lr\\_scheduler\\_type: linear\n* num\\_epochs: 10", "### Training results", "### Framework versions\n\n\n* Transformers 4.30.2\n* Pytorch 1.13.1+cu117\n* Datasets 2.13.2\n* Tokenizers 0.13.3" ]	[ "TAGS\n#transformers #pytorch #roberta #fill-mask #generated_from_trainer #license-mit #autotrain_compatible #endpoints_compatible #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 0.0001\n* train\\_batch\\_size: 32\n* eval\\_batch\\_size: 64\n* seed: 42\n* gradient\\_accumulation\\_steps: 32\n* total\\_train\\_batch\\_size: 1024\n* optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-06\n* lr\\_scheduler\\_type: linear\n* num\\_epochs: 10", "### Training results", "### Framework versions\n\n\n* Transformers 4.30.2\n* Pytorch 1.13.1+cu117\n* Datasets 2.13.2\n* Tokenizers 0.13.3" ]
text-generation	transformers	# merge This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit). ## Merge Details ### Merge Method This model was merged using the [linear](https://arxiv.org/abs/2203.05482) merge method. ### Models Merged The following models were included in the merge: * [deepseek-ai/deepseek-math-7b-rl](https://huggingface.co/deepseek-ai/deepseek-math-7b-rl) ### Configuration The following YAML configuration was used to produce this model: ```yaml merge_method: linear parameters: weight: 1.0 slices: - sources: - model: deepseek-ai/deepseek-math-7b-rl layer_range: [1, 10] - sources: - model: deepseek-ai/deepseek-math-7b-rl layer_range: [5, 15] - sources: - model: deepseek-ai/deepseek-math-7b-rl layer_range: [10, 20] - sources: - model: deepseek-ai/deepseek-math-7b-rl layer_range: [15, 25] - sources: - model: deepseek-ai/deepseek-math-7b-rl layer_range: [20, 30] dtype: float16 ```	{"library_name": "transformers", "tags": ["mergekit", "merge"], "base_model": ["deepseek-ai/deepseek-math-7b-rl"]}	Thermostatic/deepseek-math-10.8b-rl	null	[ "transformers", "safetensors", "llama", "text-generation", "mergekit", "merge", "conversational", "arxiv:2203.05482", "base_model:deepseek-ai/deepseek-math-7b-rl", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T03:23:33+00:00	[ "2203.05482" ]	[]	TAGS #transformers #safetensors #llama #text-generation #mergekit #merge #conversational #arxiv-2203.05482 #base_model-deepseek-ai/deepseek-math-7b-rl #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	# merge This is a merge of pre-trained language models created using mergekit. ## Merge Details ### Merge Method This model was merged using the linear merge method. ### Models Merged The following models were included in the merge: * deepseek-ai/deepseek-math-7b-rl ### Configuration The following YAML configuration was used to produce this model:	[ "# merge\n\nThis is a merge of pre-trained language models created using mergekit.", "## Merge Details", "### Merge Method\n\nThis model was merged using the linear merge method.", "### Models Merged\n\nThe following models were included in the merge:\n* deepseek-ai/deepseek-math-7b-rl", "### Configuration\n\nThe following YAML configuration was used to produce this model:" ]	[ "TAGS\n#transformers #safetensors #llama #text-generation #mergekit #merge #conversational #arxiv-2203.05482 #base_model-deepseek-ai/deepseek-math-7b-rl #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "# merge\n\nThis is a merge of pre-trained language models created using mergekit.", "## Merge Details", "### Merge Method\n\nThis model was merged using the linear merge method.", "### Models Merged\n\nThe following models were included in the merge:\n* deepseek-ai/deepseek-math-7b-rl", "### Configuration\n\nThe following YAML configuration was used to produce this model:" ]
fill-mask	transformers	<!-- This model card has been generated automatically according to the information Keras had access to. You should probably proofread and complete it, then remove this comment. --> # jfforero/distilbert-base-uncased-BERT-POET3 This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset. It achieves the following results on the evaluation set: - Train Loss: 5.1503 - Validation Loss: 5.1975 - Epoch: 0 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'module': 'transformers.optimization_tf', 'class_name': 'WarmUp', 'config': {'initial_learning_rate': 2e-05, 'decay_schedule_fn': {'module': 'keras.optimizers.schedules', 'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': -998, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered_name': None}, 'warmup_steps': 1000, 'power': 1.0, 'name': None}, 'registered_name': 'WarmUp'}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.01} - training_precision: float32 ### Training results \| Train Loss \| Validation Loss \| Epoch \| \|:----------:\|:---------------:\|:-----:\| \| 5.1503 \| 5.1975 \| 0 \| ### Framework versions - Transformers 4.38.2 - TensorFlow 2.15.0 - Datasets 2.18.0 - Tokenizers 0.15.2	{"license": "apache-2.0", "tags": ["generated_from_keras_callback"], "base_model": "distilbert-base-uncased", "model-index": [{"name": "jfforero/distilbert-base-uncased-BERT-POET3", "results": []}]}	jfforero/distilbert-base-uncased-BERT-POET3	null	[ "transformers", "tf", "distilbert", "fill-mask", "generated_from_keras_callback", "base_model:distilbert-base-uncased", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T03:23:38+00:00	[]	[]	TAGS #transformers #tf #distilbert #fill-mask #generated_from_keras_callback #base_model-distilbert-base-uncased #license-apache-2.0 #autotrain_compatible #endpoints_compatible #region-us	jfforero/distilbert-base-uncased-BERT-POET3 =========================================== This model is a fine-tuned version of distilbert-base-uncased on an unknown dataset. It achieves the following results on the evaluation set: * Train Loss: 5.1503 * Validation Loss: 5.1975 * Epoch: 0 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * optimizer: {'name': 'AdamWeightDecay', 'learning\_rate': {'module': 'transformers.optimization\_tf', 'class\_name': 'WarmUp', 'config': {'initial\_learning\_rate': 2e-05, 'decay\_schedule\_fn': {'module': 'keras.optimizers.schedules', 'class\_name': 'PolynomialDecay', 'config': {'initial\_learning\_rate': 2e-05, 'decay\_steps': -998, 'end\_learning\_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered\_name': None}, 'warmup\_steps': 1000, 'power': 1.0, 'name': None}, 'registered\_name': 'WarmUp'}, 'decay': 0.0, 'beta\_1': 0.9, 'beta\_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight\_decay\_rate': 0.01} * training\_precision: float32 ### Training results ### Framework versions * Transformers 4.38.2 * TensorFlow 2.15.0 * Datasets 2.18.0 * Tokenizers 0.15.2	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* optimizer: {'name': 'AdamWeightDecay', 'learning\\_rate': {'module': 'transformers.optimization\\_tf', 'class\\_name': 'WarmUp', 'config': {'initial\\_learning\\_rate': 2e-05, 'decay\\_schedule\\_fn': {'module': 'keras.optimizers.schedules', 'class\\_name': 'PolynomialDecay', 'config': {'initial\\_learning\\_rate': 2e-05, 'decay\\_steps': -998, 'end\\_learning\\_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered\\_name': None}, 'warmup\\_steps': 1000, 'power': 1.0, 'name': None}, 'registered\\_name': 'WarmUp'}, 'decay': 0.0, 'beta\\_1': 0.9, 'beta\\_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight\\_decay\\_rate': 0.01}\n* training\\_precision: float32", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* TensorFlow 2.15.0\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #tf #distilbert #fill-mask #generated_from_keras_callback #base_model-distilbert-base-uncased #license-apache-2.0 #autotrain_compatible #endpoints_compatible #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* optimizer: {'name': 'AdamWeightDecay', 'learning\\_rate': {'module': 'transformers.optimization\\_tf', 'class\\_name': 'WarmUp', 'config': {'initial\\_learning\\_rate': 2e-05, 'decay\\_schedule\\_fn': {'module': 'keras.optimizers.schedules', 'class\\_name': 'PolynomialDecay', 'config': {'initial\\_learning\\_rate': 2e-05, 'decay\\_steps': -998, 'end\\_learning\\_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered\\_name': None}, 'warmup\\_steps': 1000, 'power': 1.0, 'name': None}, 'registered\\_name': 'WarmUp'}, 'decay': 0.0, 'beta\\_1': 0.9, 'beta\\_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight\\_decay\\_rate': 0.01}\n* training\\_precision: float32", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* TensorFlow 2.15.0\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]
null	transformers	# Uploaded model - Developed by: jsfs11 - License: apache-2.0 - Finetuned from model : unsloth/llama-3-8b-bnb-4bit This llama model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library. [<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)	{"language": ["en"], "license": "apache-2.0", "tags": ["text-generation-inference", "transformers", "unsloth", "llama", "trl"], "base_model": "unsloth/llama-3-8b-bnb-4bit"}	jsfs11/alpaca3qloratest	null	[ "transformers", "safetensors", "text-generation-inference", "unsloth", "llama", "trl", "en", "base_model:unsloth/llama-3-8b-bnb-4bit", "license:apache-2.0", "endpoints_compatible", "region:us" ]	null	2024-04-19T03:25:34+00:00	[]	[ "en" ]	TAGS #transformers #safetensors #text-generation-inference #unsloth #llama #trl #en #base_model-unsloth/llama-3-8b-bnb-4bit #license-apache-2.0 #endpoints_compatible #region-us	# Uploaded model - Developed by: jsfs11 - License: apache-2.0 - Finetuned from model : unsloth/llama-3-8b-bnb-4bit This llama model was trained 2x faster with Unsloth and Huggingface's TRL library. <img src="URL width="200"/>	[ "# Uploaded model\n\n- Developed by: jsfs11\n- License: apache-2.0\n- Finetuned from model : unsloth/llama-3-8b-bnb-4bit\n\nThis llama model was trained 2x faster with Unsloth and Huggingface's TRL library.\n\n<img src=\"URL width=\"200\"/>" ]	[ "TAGS\n#transformers #safetensors #text-generation-inference #unsloth #llama #trl #en #base_model-unsloth/llama-3-8b-bnb-4bit #license-apache-2.0 #endpoints_compatible #region-us \n", "# Uploaded model\n\n- Developed by: jsfs11\n- License: apache-2.0\n- Finetuned from model : unsloth/llama-3-8b-bnb-4bit\n\nThis llama model was trained 2x faster with Unsloth and Huggingface's TRL library.\n\n<img src=\"URL width=\"200\"/>" ]
text2text-generation	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # bert2bert-abs-final-lr-5e-05-encmaxlen-512-decmaxlen-256 This model is a fine-tuned version of [](https://huggingface.co/) on the id_liputan6 dataset. It achieves the following results on the evaluation set: - Loss: 2.876 - R1 Precision: 0.2907 - R1 Recall: 0.3519 - R1 Fmeasure: 0.3149 - R2 Precision: 0.1166 - R2 Recall: 0.1408 - R2 Fmeasure: 0.1261 - Rl Precision: 0.2301 - Rl Recall: 0.2787 - Rl Fmeasure: 0.2493 ## Model description Encoder Decoder Model using IndoBERT (indobert-base-uncased) as Encoder and Decoder ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure - Max length generation: 80 - Min length generation: 10 ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 18 - eval_batch_size: 18 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 8 - mixed_precision_training: Native AMP ### Framework versions - Transformers 4.38.2 - Pytorch 2.2.1+cu121 - Datasets 2.18.0 - Tokenizers 0.15.2	{"tags": ["generated_from_trainer"], "datasets": ["id_liputan6"], "model-index": [{"name": "bert2bert-abs-final-lr-5e-05-encmaxlen-512-decmaxlen-256", "results": []}]}	Alfahluzi/bert2bert-model3	null	[ "transformers", "tensorboard", "safetensors", "encoder-decoder", "text2text-generation", "generated_from_trainer", "dataset:id_liputan6", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T03:26:27+00:00	[]	[]	TAGS #transformers #tensorboard #safetensors #encoder-decoder #text2text-generation #generated_from_trainer #dataset-id_liputan6 #autotrain_compatible #endpoints_compatible #region-us	# bert2bert-abs-final-lr-5e-05-encmaxlen-512-decmaxlen-256 This model is a fine-tuned version of [](URL on the id_liputan6 dataset. It achieves the following results on the evaluation set: - Loss: 2.876 - R1 Precision: 0.2907 - R1 Recall: 0.3519 - R1 Fmeasure: 0.3149 - R2 Precision: 0.1166 - R2 Recall: 0.1408 - R2 Fmeasure: 0.1261 - Rl Precision: 0.2301 - Rl Recall: 0.2787 - Rl Fmeasure: 0.2493 ## Model description Encoder Decoder Model using IndoBERT (indobert-base-uncased) as Encoder and Decoder ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure - Max length generation: 80 - Min length generation: 10 ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 18 - eval_batch_size: 18 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 8 - mixed_precision_training: Native AMP ### Framework versions - Transformers 4.38.2 - Pytorch 2.2.1+cu121 - Datasets 2.18.0 - Tokenizers 0.15.2	[ "# bert2bert-abs-final-lr-5e-05-encmaxlen-512-decmaxlen-256\n\nThis model is a fine-tuned version of [](URL on the id_liputan6 dataset.\nIt achieves the following results on the evaluation set:\n- Loss: 2.876\n- R1 Precision: 0.2907\n- R1 Recall: 0.3519\n- R1 Fmeasure: 0.3149\n- R2 Precision: 0.1166\n- R2 Recall: 0.1408\n- R2 Fmeasure: 0.1261\n- Rl Precision: 0.2301\n- Rl Recall: 0.2787\n- Rl Fmeasure: 0.2493", "## Model description\n\nEncoder Decoder Model using IndoBERT (indobert-base-uncased) as Encoder and Decoder", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure\n- Max length generation: 80\n- Min length generation: 10", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 5e-05\n- train_batch_size: 18\n- eval_batch_size: 18\n- seed: 42\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- num_epochs: 8\n- mixed_precision_training: Native AMP", "### Framework versions\n\n- Transformers 4.38.2\n- Pytorch 2.2.1+cu121\n- Datasets 2.18.0\n- Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #tensorboard #safetensors #encoder-decoder #text2text-generation #generated_from_trainer #dataset-id_liputan6 #autotrain_compatible #endpoints_compatible #region-us \n", "# bert2bert-abs-final-lr-5e-05-encmaxlen-512-decmaxlen-256\n\nThis model is a fine-tuned version of [](URL on the id_liputan6 dataset.\nIt achieves the following results on the evaluation set:\n- Loss: 2.876\n- R1 Precision: 0.2907\n- R1 Recall: 0.3519\n- R1 Fmeasure: 0.3149\n- R2 Precision: 0.1166\n- R2 Recall: 0.1408\n- R2 Fmeasure: 0.1261\n- Rl Precision: 0.2301\n- Rl Recall: 0.2787\n- Rl Fmeasure: 0.2493", "## Model description\n\nEncoder Decoder Model using IndoBERT (indobert-base-uncased) as Encoder and Decoder", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure\n- Max length generation: 80\n- Min length generation: 10", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 5e-05\n- train_batch_size: 18\n- eval_batch_size: 18\n- seed: 42\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- num_epochs: 8\n- mixed_precision_training: Native AMP", "### Framework versions\n\n- Transformers 4.38.2\n- Pytorch 2.2.1+cu121\n- Datasets 2.18.0\n- Tokenizers 0.15.2" ]
text-generation	transformers	# Clevyby/Silicon-Masha-7B-Q5_K_S-GGUF This model was converted to GGUF format from [`LakoMoor/Silicon-Masha-7B`](https://huggingface.co/LakoMoor/Silicon-Masha-7B) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space. Refer to the [original model card](https://huggingface.co/LakoMoor/Silicon-Masha-7B) for more details on the model. ### Note: The additional files in this GGUF repo is for personal usage using Text Gen Webui with llamacpp_hf.	{"language": ["en", "ru"], "license": "cc-by-nc-4.0", "library_name": "transformers", "tags": ["not-for-all-audiences", "nsfw", "llama-cpp", "gguf-my-repo"], "model_name": "Silicon-Masha-7B", "base_model": ["mistralai/Mistral-7B-v0.1"], "inference": false, "model_creator": "LakoMoor", "model_type": "mistral"}	Clevyby/Silicon-Masha-7B-Q5_K_S-GGUF	null	[ "transformers", "gguf", "mistral", "text-generation", "not-for-all-audiences", "nsfw", "llama-cpp", "gguf-my-repo", "en", "ru", "base_model:mistralai/Mistral-7B-v0.1", "license:cc-by-nc-4.0", "autotrain_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T03:28:47+00:00	[]	[ "en", "ru" ]	TAGS #transformers #gguf #mistral #text-generation #not-for-all-audiences #nsfw #llama-cpp #gguf-my-repo #en #ru #base_model-mistralai/Mistral-7B-v0.1 #license-cc-by-nc-4.0 #autotrain_compatible #text-generation-inference #region-us	# Clevyby/Silicon-Masha-7B-Q5_K_S-GGUF This model was converted to GGUF format from 'LakoMoor/Silicon-Masha-7B' using URL via the URL's GGUF-my-repo space. Refer to the original model card for more details on the model. ### Note: The additional files in this GGUF repo is for personal usage using Text Gen Webui with llamacpp_hf.	[ "# Clevyby/Silicon-Masha-7B-Q5_K_S-GGUF\nThis model was converted to GGUF format from 'LakoMoor/Silicon-Masha-7B' using URL via the URL's GGUF-my-repo space.\nRefer to the original model card for more details on the model.", "### Note:\nThe additional files in this GGUF repo is for personal usage using Text Gen Webui with llamacpp_hf." ]	[ "TAGS\n#transformers #gguf #mistral #text-generation #not-for-all-audiences #nsfw #llama-cpp #gguf-my-repo #en #ru #base_model-mistralai/Mistral-7B-v0.1 #license-cc-by-nc-4.0 #autotrain_compatible #text-generation-inference #region-us \n", "# Clevyby/Silicon-Masha-7B-Q5_K_S-GGUF\nThis model was converted to GGUF format from 'LakoMoor/Silicon-Masha-7B' using URL via the URL's GGUF-my-repo space.\nRefer to the original model card for more details on the model.", "### Note:\nThe additional files in this GGUF repo is for personal usage using Text Gen Webui with llamacpp_hf." ]
text-generation	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	Grayx/sad_pepe_6.0	null	[ "transformers", "safetensors", "stablelm", "text-generation", "conversational", "arxiv:1910.09700", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T03:29:14+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #stablelm #text-generation #conversational #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #stablelm #text-generation #conversational #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
text-generation	transformers	# Full Parameter Finetuning 8B 8192 context length Llama 3 on Malaysian text 9.5B tokens README at https://github.com/huseinzol05/malaya/tree/master/session/llama3 Data preparation at https://github.com/malaysia-ai/dedup-text-dataset/tree/main/llama-3 WandB, https://wandb.ai/huseinzol05/finetune-llama-3-8b/workspace?nw=nwuserhuseinzol05 ## how-to ```python from transformers import AutoTokenizer, AutoModelForCausalLM import torch tokenizer = AutoTokenizer.from_pretrained('mesolitica/llama-3-8b-8192-hf') model = AutoModelForCausalLM.from_pretrained('mesolitica/llama-3-8b-8192-hf', torch_dtype=torch.bfloat16, use_flash_attention_2 = True) _ = model.cuda() t = tokenizer('User: KWSP tu apa?\nBot: ', return_tensors = 'pt').to('cuda') r = model.generate(**t, max_new_tokens=200, top_p=0.95, top_k=50, temperature=0.9, do_sample=True, num_beams=1,) tokenizer.decode(r[0]) ``` ``` User: KWSP tu apa? Bot: ialah kumpulan wang simpanan pekerja kita sebagai badan kebajikan yang memastikan kesejahteraan 13.5 juta ahlinya terus terbela. 2) Pengeluaran i-sinar akan diteruskan dengan bayaran pertama 3) Pengeluaran i-citra akan diteruskan dengan bayaran pertama bulan April 2021. 4) 600,000 orang telah memohon untuk pengeluaran i-lestari daripada jumlah yang layak iaitu 2.5 juta orang. 5) KWSP telah meningkatkan had pengeluaran i-lestari daripada 6,000 kepada 9,000 bagi 6) Peratusan pengeluaran i-lestari turut dinaikkan daripada 4% kepada 10% iaitu maksimum pengeluaran RM ```	{"language": ["ms"]}	mesolitica/llama-3-8b-8192-hf	null	[ "transformers", "safetensors", "llama", "text-generation", "ms", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T03:32:59+00:00	[]	[ "ms" ]	TAGS #transformers #safetensors #llama #text-generation #ms #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	# Full Parameter Finetuning 8B 8192 context length Llama 3 on Malaysian text 9.5B tokens README at URL Data preparation at URL WandB, URL ## how-to	[ "# Full Parameter Finetuning 8B 8192 context length Llama 3 on Malaysian text 9.5B tokens\n\nREADME at URL\n\nData preparation at URL\n\nWandB, URL", "## how-to" ]	[ "TAGS\n#transformers #safetensors #llama #text-generation #ms #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "# Full Parameter Finetuning 8B 8192 context length Llama 3 on Malaysian text 9.5B tokens\n\nREADME at URL\n\nData preparation at URL\n\nWandB, URL", "## how-to" ]
text-classification	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # robust_llm_pythia-14m_ian-022_PasswordMatch_n-its-25 This model is a fine-tuned version of [EleutherAI/pythia-14m](https://huggingface.co/EleutherAI/pythia-14m) on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 8 - eval_batch_size: 64 - seed: 0 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results ### Framework versions - Transformers 4.39.3 - Pytorch 2.2.1 - Datasets 2.18.0 - Tokenizers 0.15.2	{"tags": ["generated_from_trainer"], "base_model": "EleutherAI/pythia-14m", "model-index": [{"name": "robust_llm_pythia-14m_ian-022_PasswordMatch_n-its-25", "results": []}]}	AlignmentResearch/robust_llm_pythia-14m_ian-022_PasswordMatch_n-its-25	null	[ "transformers", "tensorboard", "safetensors", "gpt_neox", "text-classification", "generated_from_trainer", "base_model:EleutherAI/pythia-14m", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T03:33:00+00:00	[]	[]	TAGS #transformers #tensorboard #safetensors #gpt_neox #text-classification #generated_from_trainer #base_model-EleutherAI/pythia-14m #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	# robust_llm_pythia-14m_ian-022_PasswordMatch_n-its-25 This model is a fine-tuned version of EleutherAI/pythia-14m on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 8 - eval_batch_size: 64 - seed: 0 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results ### Framework versions - Transformers 4.39.3 - Pytorch 2.2.1 - Datasets 2.18.0 - Tokenizers 0.15.2	[ "# robust_llm_pythia-14m_ian-022_PasswordMatch_n-its-25\n\nThis model is a fine-tuned version of EleutherAI/pythia-14m on an unknown dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 1e-05\n- train_batch_size: 8\n- eval_batch_size: 64\n- seed: 0\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- num_epochs: 1", "### Training results", "### Framework versions\n\n- Transformers 4.39.3\n- Pytorch 2.2.1\n- Datasets 2.18.0\n- Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #tensorboard #safetensors #gpt_neox #text-classification #generated_from_trainer #base_model-EleutherAI/pythia-14m #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "# robust_llm_pythia-14m_ian-022_PasswordMatch_n-its-25\n\nThis model is a fine-tuned version of EleutherAI/pythia-14m on an unknown dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 1e-05\n- train_batch_size: 8\n- eval_batch_size: 64\n- seed: 0\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- num_epochs: 1", "### Training results", "### Framework versions\n\n- Transformers 4.39.3\n- Pytorch 2.2.1\n- Datasets 2.18.0\n- Tokenizers 0.15.2" ]
null	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # llava_bakllava_7b_v2_8192 This model is a fine-tuned version of [llava-hf/bakLlava-v1-hf](https://huggingface.co/llava-hf/bakLlava-v1-hf) on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 1 - eval_batch_size: 1 - seed: 42 - distributed_type: multi-GPU - num_devices: 16 - gradient_accumulation_steps: 8 - total_train_batch_size: 128 - total_eval_batch_size: 16 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - lr_scheduler_warmup_ratio: 0.03 - num_epochs: 1.0 ### Training results ### Framework versions - Transformers 4.39.2 - Pytorch 2.2.1 - Datasets 2.18.0 - Tokenizers 0.15.2	{"tags": ["generated_from_trainer"], "base_model": "llava-hf/bakLlava-v1-hf", "model-index": [{"name": "llava_bakllava_7b_v2_8192", "results": []}]}	MFuyu/llava_bakllava_7b_v2_8192	null	[ "transformers", "safetensors", "llava", "pretraining", "generated_from_trainer", "base_model:llava-hf/bakLlava-v1-hf", "endpoints_compatible", "region:us" ]	null	2024-04-19T03:33:38+00:00	[]	[]	TAGS #transformers #safetensors #llava #pretraining #generated_from_trainer #base_model-llava-hf/bakLlava-v1-hf #endpoints_compatible #region-us	# llava_bakllava_7b_v2_8192 This model is a fine-tuned version of llava-hf/bakLlava-v1-hf on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 1 - eval_batch_size: 1 - seed: 42 - distributed_type: multi-GPU - num_devices: 16 - gradient_accumulation_steps: 8 - total_train_batch_size: 128 - total_eval_batch_size: 16 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - lr_scheduler_warmup_ratio: 0.03 - num_epochs: 1.0 ### Training results ### Framework versions - Transformers 4.39.2 - Pytorch 2.2.1 - Datasets 2.18.0 - Tokenizers 0.15.2	[ "# llava_bakllava_7b_v2_8192\n\nThis model is a fine-tuned version of llava-hf/bakLlava-v1-hf on an unknown dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 1e-05\n- train_batch_size: 1\n- eval_batch_size: 1\n- seed: 42\n- distributed_type: multi-GPU\n- num_devices: 16\n- gradient_accumulation_steps: 8\n- total_train_batch_size: 128\n- total_eval_batch_size: 16\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: cosine\n- lr_scheduler_warmup_ratio: 0.03\n- num_epochs: 1.0", "### Training results", "### Framework versions\n\n- Transformers 4.39.2\n- Pytorch 2.2.1\n- Datasets 2.18.0\n- Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #safetensors #llava #pretraining #generated_from_trainer #base_model-llava-hf/bakLlava-v1-hf #endpoints_compatible #region-us \n", "# llava_bakllava_7b_v2_8192\n\nThis model is a fine-tuned version of llava-hf/bakLlava-v1-hf on an unknown dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 1e-05\n- train_batch_size: 1\n- eval_batch_size: 1\n- seed: 42\n- distributed_type: multi-GPU\n- num_devices: 16\n- gradient_accumulation_steps: 8\n- total_train_batch_size: 128\n- total_eval_batch_size: 16\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: cosine\n- lr_scheduler_warmup_ratio: 0.03\n- num_epochs: 1.0", "### Training results", "### Framework versions\n\n- Transformers 4.39.2\n- Pytorch 2.2.1\n- Datasets 2.18.0\n- Tokenizers 0.15.2" ]
null	peft	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # mistralv1_spectral_r8_25e5_e2 This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the None dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.00025 - train_batch_size: 4 - eval_batch_size: 8 - seed: 0 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results ### Framework versions - PEFT 0.9.0 - Transformers 4.39.3 - Pytorch 2.2.2 - Datasets 2.18.0 - Tokenizers 0.15.2	{"library_name": "peft", "tags": ["generated_from_trainer"], "base_model": "mistralai/Mistral-7B-v0.1", "model-index": [{"name": "mistralv1_spectral_r8_25e5_e2", "results": []}]}	fangzhaoz/mistralv1_spectral_r8_25e5_e2	null	[ "peft", "safetensors", "generated_from_trainer", "base_model:mistralai/Mistral-7B-v0.1", "region:us" ]	null	2024-04-19T03:35:55+00:00	[]	[]	TAGS #peft #safetensors #generated_from_trainer #base_model-mistralai/Mistral-7B-v0.1 #region-us	# mistralv1_spectral_r8_25e5_e2 This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the None dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.00025 - train_batch_size: 4 - eval_batch_size: 8 - seed: 0 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results ### Framework versions - PEFT 0.9.0 - Transformers 4.39.3 - Pytorch 2.2.2 - Datasets 2.18.0 - Tokenizers 0.15.2	[ "# mistralv1_spectral_r8_25e5_e2\n\nThis model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the None dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 0.00025\n- train_batch_size: 4\n- eval_batch_size: 8\n- seed: 0\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- num_epochs: 2", "### Training results", "### Framework versions\n\n- PEFT 0.9.0\n- Transformers 4.39.3\n- Pytorch 2.2.2\n- Datasets 2.18.0\n- Tokenizers 0.15.2" ]	[ "TAGS\n#peft #safetensors #generated_from_trainer #base_model-mistralai/Mistral-7B-v0.1 #region-us \n", "# mistralv1_spectral_r8_25e5_e2\n\nThis model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the None dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 0.00025\n- train_batch_size: 4\n- eval_batch_size: 8\n- seed: 0\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- num_epochs: 2", "### Training results", "### Framework versions\n\n- PEFT 0.9.0\n- Transformers 4.39.3\n- Pytorch 2.2.2\n- Datasets 2.18.0\n- Tokenizers 0.15.2" ]
text-classification	transformers	<!-- This model card has been generated automatically according to the information Keras had access to. You should probably proofread and complete it, then remove this comment. --> # my-test-model This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set: ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - optimizer: None - training_precision: float32 ### Training results ### Framework versions - Transformers 4.40.0 - TensorFlow 2.15.0 - Datasets 2.18.0 - Tokenizers 0.19.1	{"tags": ["generated_from_keras_callback"], "model-index": [{"name": "my-test-model", "results": []}]}	johnwu0113/my-test-model	null	[ "transformers", "tf", "safetensors", "bert", "text-classification", "generated_from_keras_callback", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T03:38:18+00:00	[]	[]	TAGS #transformers #tf #safetensors #bert #text-classification #generated_from_keras_callback #autotrain_compatible #endpoints_compatible #region-us	# my-test-model This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set: ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - optimizer: None - training_precision: float32 ### Training results ### Framework versions - Transformers 4.40.0 - TensorFlow 2.15.0 - Datasets 2.18.0 - Tokenizers 0.19.1	[ "# my-test-model\n\nThis model was trained from scratch on an unknown dataset.\nIt achieves the following results on the evaluation set:", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- optimizer: None\n- training_precision: float32", "### Training results", "### Framework versions\n\n- Transformers 4.40.0\n- TensorFlow 2.15.0\n- Datasets 2.18.0\n- Tokenizers 0.19.1" ]	[ "TAGS\n#transformers #tf #safetensors #bert #text-classification #generated_from_keras_callback #autotrain_compatible #endpoints_compatible #region-us \n", "# my-test-model\n\nThis model was trained from scratch on an unknown dataset.\nIt achieves the following results on the evaluation set:", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- optimizer: None\n- training_precision: float32", "### Training results", "### Framework versions\n\n- Transformers 4.40.0\n- TensorFlow 2.15.0\n- Datasets 2.18.0\n- Tokenizers 0.19.1" ]
text-generation	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	fangzhaoz/mistralv1_spectral_r8_25e5_e2_merged	null	[ "transformers", "safetensors", "mistral", "text-generation", "arxiv:1910.09700", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T03:40:44+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #mistral #text-generation #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #mistral #text-generation #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
text-classification	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # robust_llm_pythia-70m_ian-022_PasswordMatch_n-its-25 This model is a fine-tuned version of [EleutherAI/pythia-70m](https://huggingface.co/EleutherAI/pythia-70m) on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 8 - eval_batch_size: 64 - seed: 0 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results ### Framework versions - Transformers 4.39.3 - Pytorch 2.2.1 - Datasets 2.18.0 - Tokenizers 0.15.2	{"license": "apache-2.0", "tags": ["generated_from_trainer"], "base_model": "EleutherAI/pythia-70m", "model-index": [{"name": "robust_llm_pythia-70m_ian-022_PasswordMatch_n-its-25", "results": []}]}	AlignmentResearch/robust_llm_pythia-70m_ian-022_PasswordMatch_n-its-25	null	[ "transformers", "tensorboard", "safetensors", "gpt_neox", "text-classification", "generated_from_trainer", "base_model:EleutherAI/pythia-70m", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T03:42:36+00:00	[]	[]	TAGS #transformers #tensorboard #safetensors #gpt_neox #text-classification #generated_from_trainer #base_model-EleutherAI/pythia-70m #license-apache-2.0 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	# robust_llm_pythia-70m_ian-022_PasswordMatch_n-its-25 This model is a fine-tuned version of EleutherAI/pythia-70m on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 8 - eval_batch_size: 64 - seed: 0 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results ### Framework versions - Transformers 4.39.3 - Pytorch 2.2.1 - Datasets 2.18.0 - Tokenizers 0.15.2	[ "# robust_llm_pythia-70m_ian-022_PasswordMatch_n-its-25\n\nThis model is a fine-tuned version of EleutherAI/pythia-70m on an unknown dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 1e-05\n- train_batch_size: 8\n- eval_batch_size: 64\n- seed: 0\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- num_epochs: 1", "### Training results", "### Framework versions\n\n- Transformers 4.39.3\n- Pytorch 2.2.1\n- Datasets 2.18.0\n- Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #tensorboard #safetensors #gpt_neox #text-classification #generated_from_trainer #base_model-EleutherAI/pythia-70m #license-apache-2.0 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "# robust_llm_pythia-70m_ian-022_PasswordMatch_n-its-25\n\nThis model is a fine-tuned version of EleutherAI/pythia-70m on an unknown dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 1e-05\n- train_batch_size: 8\n- eval_batch_size: 64\n- seed: 0\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- num_epochs: 1", "### Training results", "### Framework versions\n\n- Transformers 4.39.3\n- Pytorch 2.2.1\n- Datasets 2.18.0\n- Tokenizers 0.15.2" ]
fill-mask	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	emma7897/bert_one	null	[ "transformers", "safetensors", "bert", "fill-mask", "arxiv:1910.09700", "autotrain_compatible", "endpoints_compatible", "has_space", "region:us" ]	null	2024-04-19T03:46:11+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #bert #fill-mask #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #has_space #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #bert #fill-mask #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #has_space #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
null	peft	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # Sailor-1.8B-chatml This model is a fine-tuned version of [sail/Sailor-1.8B](https://huggingface.co/sail/Sailor-1.8B) on the generator dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 2 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 2 - total_train_batch_size: 4 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: constant - lr_scheduler_warmup_ratio: 0.03 - num_epochs: 1 ### Training results ### Framework versions - PEFT 0.8.2 - Transformers 4.38.2 - Pytorch 2.1.2 - Datasets 2.17.0 - Tokenizers 0.15.2	{"license": "apache-2.0", "library_name": "peft", "tags": ["trl", "sft", "generated_from_trainer"], "datasets": ["generator"], "base_model": "sail/Sailor-1.8B", "model-index": [{"name": "Sailor-1.8B-chatml", "results": []}]}	DuongTrongChi/Sailor-1.8B-chatml	null	[ "peft", "tensorboard", "safetensors", "qwen2", "trl", "sft", "generated_from_trainer", "dataset:generator", "base_model:sail/Sailor-1.8B", "license:apache-2.0", "region:us" ]	null	2024-04-19T03:46:30+00:00	[]	[]	TAGS #peft #tensorboard #safetensors #qwen2 #trl #sft #generated_from_trainer #dataset-generator #base_model-sail/Sailor-1.8B #license-apache-2.0 #region-us	# Sailor-1.8B-chatml This model is a fine-tuned version of sail/Sailor-1.8B on the generator dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 2 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 2 - total_train_batch_size: 4 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: constant - lr_scheduler_warmup_ratio: 0.03 - num_epochs: 1 ### Training results ### Framework versions - PEFT 0.8.2 - Transformers 4.38.2 - Pytorch 2.1.2 - Datasets 2.17.0 - Tokenizers 0.15.2	[ "# Sailor-1.8B-chatml\n\nThis model is a fine-tuned version of sail/Sailor-1.8B on the generator dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 0.0002\n- train_batch_size: 2\n- eval_batch_size: 8\n- seed: 42\n- gradient_accumulation_steps: 2\n- total_train_batch_size: 4\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: constant\n- lr_scheduler_warmup_ratio: 0.03\n- num_epochs: 1", "### Training results", "### Framework versions\n\n- PEFT 0.8.2\n- Transformers 4.38.2\n- Pytorch 2.1.2\n- Datasets 2.17.0\n- Tokenizers 0.15.2" ]	[ "TAGS\n#peft #tensorboard #safetensors #qwen2 #trl #sft #generated_from_trainer #dataset-generator #base_model-sail/Sailor-1.8B #license-apache-2.0 #region-us \n", "# Sailor-1.8B-chatml\n\nThis model is a fine-tuned version of sail/Sailor-1.8B on the generator dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 0.0002\n- train_batch_size: 2\n- eval_batch_size: 8\n- seed: 42\n- gradient_accumulation_steps: 2\n- total_train_batch_size: 4\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: constant\n- lr_scheduler_warmup_ratio: 0.03\n- num_epochs: 1", "### Training results", "### Framework versions\n\n- PEFT 0.8.2\n- Transformers 4.38.2\n- Pytorch 2.1.2\n- Datasets 2.17.0\n- Tokenizers 0.15.2" ]
null	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	PranavBP525/phi-2-storygen-v3	null	[ "transformers", "safetensors", "arxiv:1910.09700", "endpoints_compatible", "region:us" ]	null	2024-04-19T03:46:45+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #arxiv-1910.09700 #endpoints_compatible #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #arxiv-1910.09700 #endpoints_compatible #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
text-generation	null	<!-- header start --> <!-- 200823 --> <div style="width: auto; margin-left: auto; margin-right: auto"> <img src="https://github.com/LlamaEdge/LlamaEdge/raw/dev/assets/logo.svg" style="width: 100%; min-width: 400px; display: block; margin: auto;"> </div> <hr style="margin-top: 1.0em; margin-bottom: 1.0em;"> <!-- header end --> # Meta-Llama-3-8B-Instruct-GGUF ## Original Model [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) ## Run with LlamaEdge - LlamaEdge version: [v0.8.3](https://github.com/LlamaEdge/LlamaEdge/releases/tag/0.8.3) and above - Prompt template - Prompt type: `llama-3-chat` - Prompt string ```text <\|begin_of_text\|><\|start_header_id\|>system<\|end_header_id\|> {{ system_prompt }}<\|eot_id\|><\|start_header_id\|>user<\|end_header_id\|> {{ user_message_1 }}<\|eot_id\|><\|start_header_id\|>assistant<\|end_header_id\|> {{ model_answer_1 }}<\|eot_id\|><\|start_header_id\|>user<\|end_header_id\|> {{ user_message_2 }}<\|eot_id\|><\|start_header_id\|>assistant<\|end_header_id\|> ``` - Context size: `4096` - Run as LlamaEdge service ```bash wasmedge --dir .:. --nn-preload default:GGML:AUTO:Meta-Llama-3-8B-Instruct-Q5_K_M.gguf \ llama-api-server.wasm \ --prompt-template llama-3-chat \ --ctx-size 4096 \ --model-name Llama-3-8b ``` - Run as LlamaEdge command app ```bash wasmedge --dir .:. --nn-preload default:GGML:AUTO:Meta-Llama-3-8B-Instruct-Q5_K_M.gguf \ llama-chat.wasm \ --prompt-template llama-3-chat \ --ctx-size 4096 \ ``` ## Quantized GGUF Models \| Name \| Quant method \| Bits \| Size \| Use case \| \| ---- \| ---- \| ---- \| ---- \| ----- \| \| [Meta-Llama-3-8B-Instruct-Q2_K.gguf](https://huggingface.co/second-state/Llama-3-8B-Instruct-GGUF/blob/main/Meta-Llama-3-8B-Instruct-Q2_K.gguf) \| Q2_K \| 2 \| 3.18 GB\| smallest, significant quality loss - not recommended for most purposes \| \| [Meta-Llama-3-8B-Instruct-Q3_K_L.gguf](https://huggingface.co/second-state/Llama-3-8B-Instruct-GGUF/blob/main/Meta-Llama-3-8B-Instruct-Q3_K_L.gguf) \| Q3_K_L \| 3 \| 4.32 GB\| small, substantial quality loss \| \| [Meta-Llama-3-8B-Instruct-Q3_K_M.gguf](https://huggingface.co/second-state/Llama-3-8B-Instruct-GGUF/blob/main/Meta-Llama-3-8B-Instruct-Q3_K_M.gguf) \| Q3_K_M \| 3 \| 4.02 GB\| very small, high quality loss \| \| [Meta-Llama-3-8B-Instruct-Q3_K_S.gguf](https://huggingface.co/second-state/Llama-3-8B-Instruct-GGUF/blob/main/Meta-Llama-3-8B-Instruct-Q3_K_S.gguf) \| Q3_K_S \| 3 \| 3.66 GB\| very small, high quality loss \| \| [Meta-Llama-3-8B-Instruct-Q4_0.gguf](https://huggingface.co/second-state/Llama-3-8B-Instruct-GGUF/blob/main/Meta-Llama-3-8B-Instruct-Q4_0.gguf) \| Q4_0 \| 4 \| 4.66 GB\| legacy; small, very high quality loss - prefer using Q3_K_M \| \| [Meta-Llama-3-8B-Instruct-Q4_K_M.gguf](https://huggingface.co/second-state/Llama-3-8B-Instruct-GGUF/blob/main/Meta-Llama-3-8B-Instruct-Q4_K_M.gguf) \| Q4_K_M \| 4 \| 4.92 GB\| medium, balanced quality - recommended \| \| [Meta-Llama-3-8B-Instruct-Q4_K_S.gguf](https://huggingface.co/second-state/Llama-3-8B-Instruct-GGUF/blob/main/Meta-Llama-3-8B-Instruct-Q4_K_S.gguf) \| Q4_K_S \| 4 \| 4.69 GB\| small, greater quality loss \| \| [Meta-Llama-3-8B-Instruct-Q5_0.gguf](https://huggingface.co/second-state/Llama-3-8B-Instruct-GGUF/blob/main/Meta-Llama-3-8B-Instruct-Q5_0.gguf) \| Q5_0 \| 5 \| 5.6 GB\| legacy; medium, balanced quality - prefer using Q4_K_M \| \| [Meta-Llama-3-8B-Instruct-Q5_K_M.gguf](https://huggingface.co/second-state/Llama-3-8B-Instruct-GGUF/blob/main/Meta-Llama-3-8B-Instruct-Q5_K_M.gguf) \| Q5_K_M \| 5 \| 5.73 GB\| large, very low quality loss - recommended \| \| [Meta-Llama-3-8B-Instruct-Q5_K_S.gguf](https://huggingface.co/second-state/Llama-3-8B-Instruct-GGUF/blob/main/Meta-Llama-3-8B-Instruct-Q5_K_S.gguf) \| Q5_K_S \| 5 \| 5.6 GB\| large, low quality loss - recommended \| \| [Meta-Llama-3-8B-Instruct-Q6_K.gguf](https://huggingface.co/second-state/Llama-3-8B-Instruct-GGUF/blob/main/Meta-Llama-3-8B-Instruct-Q6_K.gguf) \| Q6_K \| 6 \| 6.6 GB\| very large, extremely low quality loss \| \| [Meta-Llama-3-8B-Instruct-Q8_0.gguf](https://huggingface.co/second-state/Llama-3-8B-Instruct-GGUF/blob/main/Meta-Llama-3-8B-Instruct-Q8_0.gguf) \| Q8_0 \| 8 \| 8.54 GB\| very large, extremely low quality loss - not recommended \| \| [Meta-Llama-3-8B-Instruct-f16.gguf](https://huggingface.co/second-state/Llama-3-8B-Instruct-GGUF/blob/main/Meta-Llama-3-8B-Instruct-f16.gguf) \| f16 \| 16 \| 16.1 GB\| \| Quantized with llama.cpp b2715.	{"language": ["en"], "license": "other", "model_name": "Llama3 8B Instruct", "license_name": "llama3", "arxiv": 2307.09288, "base_model": "meta-llama/Meta-Llama-3-8B-Instruct", "inference": false, "model_creator": "Meta Llama3", "model_type": "llama", "pipeline_tag": "text-generation", "quantized_by": "Second State Inc."}	second-state/Llama-3-8B-Instruct-GGUF	null	[ "gguf", "text-generation", "en", "base_model:meta-llama/Meta-Llama-3-8B-Instruct", "license:other", "region:us" ]	null	2024-04-19T03:47:19+00:00	[]	[ "en" ]	TAGS #gguf #text-generation #en #base_model-meta-llama/Meta-Llama-3-8B-Instruct #license-other #region-us	![](URL style=) --- Meta-Llama-3-8B-Instruct-GGUF ============================= Original Model -------------- meta-llama/Meta-Llama-3-8B-Instruct Run with LlamaEdge ------------------ * LlamaEdge version: v0.8.3 and above * Prompt template + Prompt type: 'llama-3-chat' + Prompt string * Context size: '4096' * Run as LlamaEdge service * Run as LlamaEdge command app Quantized GGUF Models --------------------- Quantized with URL b2715.	[]	[ "TAGS\n#gguf #text-generation #en #base_model-meta-llama/Meta-Llama-3-8B-Instruct #license-other #region-us \n" ]
object-detection	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	mali17361/detr-finetuned-balloon-v1	null	[ "transformers", "safetensors", "table-transformer", "object-detection", "arxiv:1910.09700", "endpoints_compatible", "region:us" ]	null	2024-04-19T03:52:13+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #table-transformer #object-detection #arxiv-1910.09700 #endpoints_compatible #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #table-transformer #object-detection #arxiv-1910.09700 #endpoints_compatible #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
feature-extraction	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # finetuned_bge_ver21 This model is a fine-tuned version of [BAAI/bge-m3](https://huggingface.co/BAAI/bge-m3) on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0001 - train_batch_size: 16 - train_group_size: 4 - eval_batch_size: 8 - seed: 42 - distributed_type: multi-GPU - num_devices: 2 - total_train_batch_size: 32 - total_eval_batch_size: 16 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 20.0 - mixed_precision_training: Native AMP ### Training results ### Framework versions - Transformers 4.39.3 - Pytorch 2.1.2 - Datasets 2.18.0 - Tokenizers 0.15.2	{"license": "mit", "tags": ["generated_from_trainer"], "base_model": "BAAI/bge-m3", "model-index": [{"name": "finetuned_bge_ver21", "results": []}]}	comet24082002/finetuned_bge_ver21	null	[ "transformers", "tensorboard", "safetensors", "xlm-roberta", "feature-extraction", "generated_from_trainer", "base_model:BAAI/bge-m3", "license:mit", "endpoints_compatible", "region:us" ]	null	2024-04-19T03:52:36+00:00	[]	[]	TAGS #transformers #tensorboard #safetensors #xlm-roberta #feature-extraction #generated_from_trainer #base_model-BAAI/bge-m3 #license-mit #endpoints_compatible #region-us	# finetuned_bge_ver21 This model is a fine-tuned version of BAAI/bge-m3 on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0001 - train_batch_size: 16 - train_group_size: 4 - eval_batch_size: 8 - seed: 42 - distributed_type: multi-GPU - num_devices: 2 - total_train_batch_size: 32 - total_eval_batch_size: 16 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 20.0 - mixed_precision_training: Native AMP ### Training results ### Framework versions - Transformers 4.39.3 - Pytorch 2.1.2 - Datasets 2.18.0 - Tokenizers 0.15.2	[ "# finetuned_bge_ver21\n\nThis model is a fine-tuned version of BAAI/bge-m3 on an unknown dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 0.0001\n- train_batch_size: 16\n- train_group_size: 4\n- eval_batch_size: 8\n- seed: 42\n- distributed_type: multi-GPU\n- num_devices: 2\n- total_train_batch_size: 32\n- total_eval_batch_size: 16\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- lr_scheduler_warmup_ratio: 0.1\n- num_epochs: 20.0\n- mixed_precision_training: Native AMP", "### Training results", "### Framework versions\n\n- Transformers 4.39.3\n- Pytorch 2.1.2\n- Datasets 2.18.0\n- Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #tensorboard #safetensors #xlm-roberta #feature-extraction #generated_from_trainer #base_model-BAAI/bge-m3 #license-mit #endpoints_compatible #region-us \n", "# finetuned_bge_ver21\n\nThis model is a fine-tuned version of BAAI/bge-m3 on an unknown dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 0.0001\n- train_batch_size: 16\n- train_group_size: 4\n- eval_batch_size: 8\n- seed: 42\n- distributed_type: multi-GPU\n- num_devices: 2\n- total_train_batch_size: 32\n- total_eval_batch_size: 16\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- lr_scheduler_warmup_ratio: 0.1\n- num_epochs: 20.0\n- mixed_precision_training: Native AMP", "### Training results", "### Framework versions\n\n- Transformers 4.39.3\n- Pytorch 2.1.2\n- Datasets 2.18.0\n- Tokenizers 0.15.2" ]
null	null	# Introduction Models in this repo are converted from https://github.com/RicherMans/CED	{}	k2-fsa/sherpa-onnx-ced-tiny-audio-tagging-2024-04-19	null	[ "onnx", "region:us" ]	null	2024-04-19T03:53:32+00:00	[]	[]	TAGS #onnx #region-us	# Introduction Models in this repo are converted from URL	[ "# Introduction\n\nModels in this repo are converted from\nURL" ]	[ "TAGS\n#onnx #region-us \n", "# Introduction\n\nModels in this repo are converted from\nURL" ]
text-generation	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	lxsure/Sniper_32	null	[ "transformers", "safetensors", "stablelm", "text-generation", "conversational", "arxiv:1910.09700", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T03:53:40+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #stablelm #text-generation #conversational #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #stablelm #text-generation #conversational #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
null	null	# Introduction Models in this repo are converted from https://github.com/RicherMans/CED	{}	k2-fsa/sherpa-onnx-ced-mini-audio-tagging-2024-04-19	null	[ "onnx", "region:us" ]	null	2024-04-19T03:54:18+00:00	[]	[]	TAGS #onnx #region-us	# Introduction Models in this repo are converted from URL	[ "# Introduction\n\nModels in this repo are converted from\nURL" ]	[ "TAGS\n#onnx #region-us \n", "# Introduction\n\nModels in this repo are converted from\nURL" ]
null	null	# Introduction Models in this repo are converted from https://github.com/RicherMans/CED	{}	k2-fsa/sherpa-onnx-ced-small-audio-tagging-2024-04-19	null	[ "onnx", "region:us" ]	null	2024-04-19T03:54:31+00:00	[]	[]	TAGS #onnx #region-us	# Introduction Models in this repo are converted from URL	[ "# Introduction\n\nModels in this repo are converted from\nURL" ]	[ "TAGS\n#onnx #region-us \n", "# Introduction\n\nModels in this repo are converted from\nURL" ]
null	null	# Introduction Models in this repo are converted from https://github.com/RicherMans/CED	{}	k2-fsa/sherpa-onnx-ced-base-audio-tagging-2024-04-19	null	[ "onnx", "region:us" ]	null	2024-04-19T03:54:41+00:00	[]	[]	TAGS #onnx #region-us	# Introduction Models in this repo are converted from URL	[ "# Introduction\n\nModels in this repo are converted from\nURL" ]	[ "TAGS\n#onnx #region-us \n", "# Introduction\n\nModels in this repo are converted from\nURL" ]
automatic-speech-recognition	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # fineturning-with-pretraining-3 This model is a fine-tuned version of [Aviral2412/mini_model](https://huggingface.co/Aviral2412/mini_model) on the common_voice_1_0 dataset. It achieves the following results on the evaluation set: - Loss: 2.9160 - Wer: 1.0000 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0001 - train_batch_size: 32 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 1000 - num_epochs: 30 - mixed_precision_training: Native AMP ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Wer \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:------:\| \| 6.0042 \| 4.27 \| 500 \| 3.1928 \| 1.0000 \| \| 2.9673 \| 8.55 \| 1000 \| 3.0856 \| 1.0000 \| \| 2.9929 \| 12.82 \| 1500 \| 3.0173 \| 1.0000 \| \| 2.9458 \| 17.09 \| 2000 \| 2.9282 \| 1.0000 \| \| 2.9084 \| 21.37 \| 2500 \| 2.9734 \| 1.0000 \| \| 2.8651 \| 25.64 \| 3000 \| 2.9234 \| 1.0000 \| \| 2.8307 \| 29.91 \| 3500 \| 2.9160 \| 1.0000 \| ### Framework versions - Transformers 4.39.3 - Pytorch 2.1.2 - Datasets 2.18.0 - Tokenizers 0.15.2	{"tags": ["generated_from_trainer"], "datasets": ["common_voice_1_0"], "metrics": ["wer"], "base_model": "Aviral2412/mini_model", "model-index": [{"name": "fineturning-with-pretraining-3", "results": [{"task": {"type": "automatic-speech-recognition", "name": "Automatic Speech Recognition"}, "dataset": {"name": "common_voice_1_0", "type": "common_voice_1_0", "config": "en", "split": "validation", "args": "en"}, "metrics": [{"type": "wer", "value": 1.0000323289796973, "name": "Wer"}]}]}]}	Aviral2412/fineturning-with-pretraining-3	null	[ "transformers", "tensorboard", "safetensors", "wav2vec2", "automatic-speech-recognition", "generated_from_trainer", "dataset:common_voice_1_0", "base_model:Aviral2412/mini_model", "model-index", "endpoints_compatible", "region:us" ]	null	2024-04-19T03:54:52+00:00	[]	[]	TAGS #transformers #tensorboard #safetensors #wav2vec2 #automatic-speech-recognition #generated_from_trainer #dataset-common_voice_1_0 #base_model-Aviral2412/mini_model #model-index #endpoints_compatible #region-us	fineturning-with-pretraining-3 ============================== This model is a fine-tuned version of Aviral2412/mini\_model on the common\_voice\_1\_0 dataset. It achieves the following results on the evaluation set: * Loss: 2.9160 * Wer: 1.0000 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * learning\_rate: 0.0001 * train\_batch\_size: 32 * eval\_batch\_size: 8 * seed: 42 * optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 * lr\_scheduler\_type: linear * lr\_scheduler\_warmup\_steps: 1000 * num\_epochs: 30 * mixed\_precision\_training: Native AMP ### Training results ### Framework versions * Transformers 4.39.3 * Pytorch 2.1.2 * Datasets 2.18.0 * Tokenizers 0.15.2	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 0.0001\n* train\\_batch\\_size: 32\n* eval\\_batch\\_size: 8\n* seed: 42\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* lr\\_scheduler\\_warmup\\_steps: 1000\n* num\\_epochs: 30\n* mixed\\_precision\\_training: Native AMP", "### Training results", "### Framework versions\n\n\n* Transformers 4.39.3\n* Pytorch 2.1.2\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #tensorboard #safetensors #wav2vec2 #automatic-speech-recognition #generated_from_trainer #dataset-common_voice_1_0 #base_model-Aviral2412/mini_model #model-index #endpoints_compatible #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 0.0001\n* train\\_batch\\_size: 32\n* eval\\_batch\\_size: 8\n* seed: 42\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* lr\\_scheduler\\_warmup\\_steps: 1000\n* num\\_epochs: 30\n* mixed\\_precision\\_training: Native AMP", "### Training results", "### Framework versions\n\n\n* Transformers 4.39.3\n* Pytorch 2.1.2\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]
fill-mask	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	emma7897/distilbert_one	null	[ "transformers", "safetensors", "distilbert", "fill-mask", "arxiv:1910.09700", "autotrain_compatible", "endpoints_compatible", "has_space", "region:us" ]	null	2024-04-19T03:57:03+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #distilbert #fill-mask #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #has_space #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #distilbert #fill-mask #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #has_space #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
text-generation	null	# Nimue 8B There is a new training script for this release. The responses are shorter in the "improved" datasets. ## Prompt format The model was trained on a zero-shot Alpaca instruction format: ``` Below is an instruction that describes a task. Write a response that appropriately completes the request. ### Instruction: {system prompt} ### Input: User: Wait a minute. Assistant: Assistant's heart skipped a beat, she hadn't expected to meet anyone today. User: Hey, didn't I see you at the library yesterday? Traits: Shy Length: Short ### Response: ``` After several attempts, I have decided not to support multi-turn conversation for the time being. You can use labels (traits, length) to control the assistant's behavior before the response field. ## Datasets Datasets about unexpected events: - allenai/UNcommonsense (conversation format) - grimulkan/theory-of-mind (summarization) - twodgirl/tama (a cat talks to its owner) Datasets about personality traits: - allenai/soda - IlyaGusev/pippa_scored - twodgirl/ewheel - twodgirl/pi (conversation made up by Pi, the emotionally intelligent chatbot) Datasets by response length: - athirdpath/Roleplay-Alpaca-NSFW (long) - fnlp/character-llm-data (short) - twodgirl/kimiko_v3 (short) - twodgirl/theory-of-mind (short summarization) - twodgirl/pi (short) ## Personality traits There are more than 100 of them in the datasets. Affectionate, Afraid, Aggressive, Alarmed, Alert, Ambitious, Amiable, Amorous, Amused, Angry, Annoyed, Anxious, Apathetic, Apologetic, Argumentative, Aroused, Arrogant, Ashamed, Assertive, Astonished, Attentive, Bellicosity, Bitter, Bluntness, Bored, Calm, Capriciousness, Caring, Cautious, Compassionate, Competitive, Concerned, Confident, Confused, Content, Courageous, Creative, Critical, Cruelty, Curious, Defiant, Depressed, Desperate, Despondent, Determined, Disappointed, Disgusted, Disobedient, Dissatisfied, Doubtful, Efficient, Embarrassed, Empathetic, Encouraging, Enthusiastic, Envious, Excited, Exhausted, Expectant, Fidelity, Forgetful, Forgiving, Fragility, Friendly, Frugal, Frustrated, Generous, Grateful, Guilty, Happy, Hateful, Helpful, Helpless, Hesitant, Homesick, Honest, Hopeful, Hostile, Impatient, Impulsive, Indecisive, Indignant, Insecure, Insulted, Integrity, Interested, Jealous, Joyous, Kind, Kindness, Loathing, Longing, Loquacity, Lost, Loving, Loyal, Lusting, Miserable, Motivated, Nervous, Nostalgic, Optimistic, Organized, Passionate, Patient, Pensive, Persistent, Persuasive, Playful, Pleased, Polite, Protective, Proud, Rebellious, Relaxed, Relieved, Remorseful, Resilient, Restless, Reverent, Sad, Scared, Self-critical, Selfish, Sentimental, Serene, Serious, Shy, Shyness, Sleepy, Startled, Stubbornness, Superior, Supportive, Suspicious, Sympathetic, Tender, Tense, Thoughtful, Tired, Understanding, Upset, Wisdom, Worried. ## References Scherer KR. What are emotions? And how can they be measured? MIT An Affective Model of Interplay Between Emotions and Learning Scherer KR. The GRID meets the wheel Manshad Abbasi Mohsin Summarizing Emotions from Text Using Plutchik’s Wheel of Emotions	{"language": ["en"], "license": "other", "tags": ["causal-lm", "llama-3"], "datasets": ["athirdpath/DPO_Pairs-Roleplay-Alpaca-NSFW", "allenai/UNcommonsense", "ClericalAid/roleplay-scripts", "fnlp/character-llm-data", "IlyaGusev/pippa_scored"], "pipeline_tag": "text-generation", "license_name": "llama3", "license_link": "LICENSE", "base_model": "meta-llama/Meta-Llama-3-8B-Instruct"}	twodgirl/Nimue-8B	null	[ "safetensors", "causal-lm", "llama-3", "text-generation", "en", "dataset:athirdpath/DPO_Pairs-Roleplay-Alpaca-NSFW", "dataset:allenai/UNcommonsense", "dataset:ClericalAid/roleplay-scripts", "dataset:fnlp/character-llm-data", "dataset:IlyaGusev/pippa_scored", "base_model:meta-llama/Meta-Llama-3-8B-Instruct", "license:other", "region:us" ]	null	2024-04-19T03:58:38+00:00	[]	[ "en" ]	TAGS #safetensors #causal-lm #llama-3 #text-generation #en #dataset-athirdpath/DPO_Pairs-Roleplay-Alpaca-NSFW #dataset-allenai/UNcommonsense #dataset-ClericalAid/roleplay-scripts #dataset-fnlp/character-llm-data #dataset-IlyaGusev/pippa_scored #base_model-meta-llama/Meta-Llama-3-8B-Instruct #license-other #region-us	# Nimue 8B There is a new training script for this release. The responses are shorter in the "improved" datasets. ## Prompt format The model was trained on a zero-shot Alpaca instruction format: After several attempts, I have decided not to support multi-turn conversation for the time being. You can use labels (traits, length) to control the assistant's behavior before the response field. ## Datasets Datasets about unexpected events: - allenai/UNcommonsense (conversation format) - grimulkan/theory-of-mind (summarization) - twodgirl/tama (a cat talks to its owner) Datasets about personality traits: - allenai/soda - IlyaGusev/pippa_scored - twodgirl/ewheel - twodgirl/pi (conversation made up by Pi, the emotionally intelligent chatbot) Datasets by response length: - athirdpath/Roleplay-Alpaca-NSFW (long) - fnlp/character-llm-data (short) - twodgirl/kimiko_v3 (short) - twodgirl/theory-of-mind (short summarization) - twodgirl/pi (short) ## Personality traits There are more than 100 of them in the datasets. Affectionate, Afraid, Aggressive, Alarmed, Alert, Ambitious, Amiable, Amorous, Amused, Angry, Annoyed, Anxious, Apathetic, Apologetic, Argumentative, Aroused, Arrogant, Ashamed, Assertive, Astonished, Attentive, Bellicosity, Bitter, Bluntness, Bored, Calm, Capriciousness, Caring, Cautious, Compassionate, Competitive, Concerned, Confident, Confused, Content, Courageous, Creative, Critical, Cruelty, Curious, Defiant, Depressed, Desperate, Despondent, Determined, Disappointed, Disgusted, Disobedient, Dissatisfied, Doubtful, Efficient, Embarrassed, Empathetic, Encouraging, Enthusiastic, Envious, Excited, Exhausted, Expectant, Fidelity, Forgetful, Forgiving, Fragility, Friendly, Frugal, Frustrated, Generous, Grateful, Guilty, Happy, Hateful, Helpful, Helpless, Hesitant, Homesick, Honest, Hopeful, Hostile, Impatient, Impulsive, Indecisive, Indignant, Insecure, Insulted, Integrity, Interested, Jealous, Joyous, Kind, Kindness, Loathing, Longing, Loquacity, Lost, Loving, Loyal, Lusting, Miserable, Motivated, Nervous, Nostalgic, Optimistic, Organized, Passionate, Patient, Pensive, Persistent, Persuasive, Playful, Pleased, Polite, Protective, Proud, Rebellious, Relaxed, Relieved, Remorseful, Resilient, Restless, Reverent, Sad, Scared, Self-critical, Selfish, Sentimental, Serene, Serious, Shy, Shyness, Sleepy, Startled, Stubbornness, Superior, Supportive, Suspicious, Sympathetic, Tender, Tense, Thoughtful, Tired, Understanding, Upset, Wisdom, Worried. ## References Scherer KR. What are emotions? And how can they be measured? MIT An Affective Model of Interplay Between Emotions and Learning Scherer KR. The GRID meets the wheel Manshad Abbasi Mohsin Summarizing Emotions from Text Using Plutchik’s Wheel of Emotions	[ "# Nimue 8B\n\nThere is a new training script for this release.\nThe responses are shorter in the \"improved\" datasets.", "## Prompt format\n\nThe model was trained on a zero-shot Alpaca instruction format:\n\n\n\nAfter several attempts, I have decided not to support multi-turn conversation for the time being. You can use labels (traits, length) to control the assistant's behavior before the response field.", "## Datasets\n\nDatasets about unexpected events:\n- allenai/UNcommonsense (conversation format)\n- grimulkan/theory-of-mind (summarization)\n- twodgirl/tama (a cat talks to its owner)\n\nDatasets about personality traits:\n- allenai/soda\n- IlyaGusev/pippa_scored\n- twodgirl/ewheel\n- twodgirl/pi (conversation made up by Pi, the emotionally intelligent chatbot)\n\nDatasets by response length:\n- athirdpath/Roleplay-Alpaca-NSFW (long)\n- fnlp/character-llm-data (short)\n- twodgirl/kimiko_v3 (short)\n- twodgirl/theory-of-mind (short summarization)\n- twodgirl/pi (short)", "## Personality traits\n\nThere are more than 100 of them in the datasets.\n\nAffectionate, Afraid, Aggressive, Alarmed, Alert, Ambitious, Amiable, Amorous, Amused, Angry, Annoyed, Anxious, Apathetic, Apologetic, Argumentative, Aroused, Arrogant, Ashamed, Assertive, Astonished, Attentive, Bellicosity, Bitter, Bluntness, Bored, Calm, Capriciousness, Caring, Cautious, Compassionate, Competitive, Concerned, Confident, Confused, Content, Courageous, Creative, Critical, Cruelty, Curious, Defiant, Depressed, Desperate, Despondent, Determined, Disappointed, Disgusted, Disobedient, Dissatisfied, Doubtful, Efficient, Embarrassed, Empathetic, Encouraging, Enthusiastic, Envious, Excited, Exhausted, Expectant, Fidelity, Forgetful, Forgiving, Fragility, Friendly, Frugal, Frustrated, Generous, Grateful, Guilty, Happy, Hateful, Helpful, Helpless, Hesitant, Homesick, Honest, Hopeful, Hostile, Impatient, Impulsive, Indecisive, Indignant, Insecure, Insulted, Integrity, Interested, Jealous, Joyous, Kind, Kindness, Loathing, Longing, Loquacity, Lost, Loving, Loyal, Lusting, Miserable, Motivated, Nervous, Nostalgic, Optimistic, Organized, Passionate, Patient, Pensive, Persistent, Persuasive, Playful, Pleased, Polite, Protective, Proud, Rebellious, Relaxed, Relieved, Remorseful, Resilient, Restless, Reverent, Sad, Scared, Self-critical, Selfish, Sentimental, Serene, Serious, Shy, Shyness, Sleepy, Startled, Stubbornness, Superior, Supportive, Suspicious, Sympathetic, Tender, Tense, Thoughtful, Tired, Understanding, Upset, Wisdom, Worried.", "## References\n\nScherer KR. What are emotions? And how can they be measured?\n\nMIT An Affective Model of Interplay Between Emotions and Learning\n\nScherer KR. The GRID meets the wheel\n\nManshad Abbasi Mohsin Summarizing Emotions from Text Using Plutchik’s Wheel of Emotions" ]	[ "TAGS\n#safetensors #causal-lm #llama-3 #text-generation #en #dataset-athirdpath/DPO_Pairs-Roleplay-Alpaca-NSFW #dataset-allenai/UNcommonsense #dataset-ClericalAid/roleplay-scripts #dataset-fnlp/character-llm-data #dataset-IlyaGusev/pippa_scored #base_model-meta-llama/Meta-Llama-3-8B-Instruct #license-other #region-us \n", "# Nimue 8B\n\nThere is a new training script for this release.\nThe responses are shorter in the \"improved\" datasets.", "## Prompt format\n\nThe model was trained on a zero-shot Alpaca instruction format:\n\n\n\nAfter several attempts, I have decided not to support multi-turn conversation for the time being. You can use labels (traits, length) to control the assistant's behavior before the response field.", "## Datasets\n\nDatasets about unexpected events:\n- allenai/UNcommonsense (conversation format)\n- grimulkan/theory-of-mind (summarization)\n- twodgirl/tama (a cat talks to its owner)\n\nDatasets about personality traits:\n- allenai/soda\n- IlyaGusev/pippa_scored\n- twodgirl/ewheel\n- twodgirl/pi (conversation made up by Pi, the emotionally intelligent chatbot)\n\nDatasets by response length:\n- athirdpath/Roleplay-Alpaca-NSFW (long)\n- fnlp/character-llm-data (short)\n- twodgirl/kimiko_v3 (short)\n- twodgirl/theory-of-mind (short summarization)\n- twodgirl/pi (short)", "## Personality traits\n\nThere are more than 100 of them in the datasets.\n\nAffectionate, Afraid, Aggressive, Alarmed, Alert, Ambitious, Amiable, Amorous, Amused, Angry, Annoyed, Anxious, Apathetic, Apologetic, Argumentative, Aroused, Arrogant, Ashamed, Assertive, Astonished, Attentive, Bellicosity, Bitter, Bluntness, Bored, Calm, Capriciousness, Caring, Cautious, Compassionate, Competitive, Concerned, Confident, Confused, Content, Courageous, Creative, Critical, Cruelty, Curious, Defiant, Depressed, Desperate, Despondent, Determined, Disappointed, Disgusted, Disobedient, Dissatisfied, Doubtful, Efficient, Embarrassed, Empathetic, Encouraging, Enthusiastic, Envious, Excited, Exhausted, Expectant, Fidelity, Forgetful, Forgiving, Fragility, Friendly, Frugal, Frustrated, Generous, Grateful, Guilty, Happy, Hateful, Helpful, Helpless, Hesitant, Homesick, Honest, Hopeful, Hostile, Impatient, Impulsive, Indecisive, Indignant, Insecure, Insulted, Integrity, Interested, Jealous, Joyous, Kind, Kindness, Loathing, Longing, Loquacity, Lost, Loving, Loyal, Lusting, Miserable, Motivated, Nervous, Nostalgic, Optimistic, Organized, Passionate, Patient, Pensive, Persistent, Persuasive, Playful, Pleased, Polite, Protective, Proud, Rebellious, Relaxed, Relieved, Remorseful, Resilient, Restless, Reverent, Sad, Scared, Self-critical, Selfish, Sentimental, Serene, Serious, Shy, Shyness, Sleepy, Startled, Stubbornness, Superior, Supportive, Suspicious, Sympathetic, Tender, Tense, Thoughtful, Tired, Understanding, Upset, Wisdom, Worried.", "## References\n\nScherer KR. What are emotions? And how can they be measured?\n\nMIT An Affective Model of Interplay Between Emotions and Learning\n\nScherer KR. The GRID meets the wheel\n\nManshad Abbasi Mohsin Summarizing Emotions from Text Using Plutchik’s Wheel of Emotions" ]
text-generation	transformers	AI Model Name: Llama 3 8B "Built with Meta Llama 3" https://llama.meta.com/llama3/license/ This is the result of running AutoAWQ to quantize the LLaMA-3 8B model to ~4 bits/parameter. To launch an OpenAI-compatible API endpoint on your Linux server: ``` git lfs install git clone https://huggingface.co/catid/cat-llama-3-8b-awq-q128-w4-gemm conda create -n vllm8 python=3.10 -y && conda activate vllm8 pip install -U git+https://github.com/vllm-project/vllm.git@a134ef6 python -m vllm.entrypoints.openai.api_server --model cat-llama-3-8b-awq-q128-w4-gemm ``` To use 2 GPUs add `--tensor-parallel-size 2 --gpu-memory-utilization 0.95`: ``` python -m vllm.entrypoints.openai.api_server --model cat-llama-3-8b-awq-q128-w4-gemm --tensor-parallel-size 2 --gpu-memory-utilization 0.95 ``` My personal TextWorld common-sense reasoning benchmark ( https://github.com/catid/textworld_llm_benchmark ) results for this model: ``` cat-llama-3-8b-awq-q128-w4-gemm : Average Score: 2.02 ± 0.29 Mixtral 8x7B : Average Score: 2.22 ± 0.33 GPT 3.5 : Average Score: 2.8 ± 1.69 ``` This is very respectable for a relatively small model!	{}	catid/cat-llama-3-8b-awq-q128-w4-gemm	null	[ "transformers", "safetensors", "llama", "text-generation", "conversational", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "4-bit", "region:us" ]	null	2024-04-19T04:02:20+00:00	[]	[]	TAGS #transformers #safetensors #llama #text-generation #conversational #autotrain_compatible #endpoints_compatible #text-generation-inference #4-bit #region-us	AI Model Name: Llama 3 8B "Built with Meta Llama 3" URL This is the result of running AutoAWQ to quantize the LLaMA-3 8B model to ~4 bits/parameter. To launch an OpenAI-compatible API endpoint on your Linux server: To use 2 GPUs add '--tensor-parallel-size 2 --gpu-memory-utilization 0.95': My personal TextWorld common-sense reasoning benchmark ( URL ) results for this model: This is very respectable for a relatively small model!	[]	[ "TAGS\n#transformers #safetensors #llama #text-generation #conversational #autotrain_compatible #endpoints_compatible #text-generation-inference #4-bit #region-us \n" ]
text2text-generation	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # flant-t5-base-function-calling-v2 This model is a fine-tuned version of [google/flan-t5-base](https://huggingface.co/google/flan-t5-base) on a Custom made Function Calling dataset. It achieves the following results on the evaluation set: - Loss: 0.0000 - Rouge1: 58.4419 - Rouge2: 51.6603 - Rougel: 58.4402 - Rougelsum: 58.4419 - Gen Len: 19.0 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 3 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Rouge1 \| Rouge2 \| Rougel \| Rougelsum \| Gen Len \| \|:-------------:\|:-----:\|:-----:\|:---------------:\|:-------:\|:-------:\|:-------:\|:---------:\|:-------:\| \| 0.0 \| 1.0 \| 15938 \| 0.0000 \| 58.4419 \| 51.6602 \| 58.44 \| 58.4418 \| 19.0 \| \| 0.0 \| 2.0 \| 31876 \| 0.0000 \| 58.4419 \| 51.6603 \| 58.4402 \| 58.4419 \| 19.0 \| \| 0.0 \| 3.0 \| 47814 \| 0.0000 \| 58.4419 \| 51.6603 \| 58.4402 \| 58.4419 \| 19.0 \| ### Framework versions - Transformers 4.38.2 - Pytorch 2.2.1+cu121 - Datasets 2.18.0 - Tokenizers 0.15.2	{"license": "apache-2.0", "tags": ["generated_from_trainer"], "metrics": ["rouge"], "base_model": "google/flan-t5-base", "model-index": [{"name": "flant-t5-base-function-calling-v2", "results": []}]}	jrcastropy/flan-t5-base-query-extraction-v2	null	[ "transformers", "tensorboard", "safetensors", "t5", "text2text-generation", "generated_from_trainer", "base_model:google/flan-t5-base", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T04:03:52+00:00	[]	[]	TAGS #transformers #tensorboard #safetensors #t5 #text2text-generation #generated_from_trainer #base_model-google/flan-t5-base #license-apache-2.0 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	flant-t5-base-function-calling-v2 ================================= This model is a fine-tuned version of google/flan-t5-base on a Custom made Function Calling dataset. It achieves the following results on the evaluation set: * Loss: 0.0000 * Rouge1: 58.4419 * Rouge2: 51.6603 * Rougel: 58.4402 * Rougelsum: 58.4419 * Gen Len: 19.0 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * learning\_rate: 5e-05 * train\_batch\_size: 8 * eval\_batch\_size: 8 * seed: 42 * optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 * lr\_scheduler\_type: linear * num\_epochs: 3 ### Training results ### Framework versions * Transformers 4.38.2 * Pytorch 2.2.1+cu121 * Datasets 2.18.0 * Tokenizers 0.15.2	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 5e-05\n* train\\_batch\\_size: 8\n* eval\\_batch\\_size: 8\n* seed: 42\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* num\\_epochs: 3", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* Pytorch 2.2.1+cu121\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #tensorboard #safetensors #t5 #text2text-generation #generated_from_trainer #base_model-google/flan-t5-base #license-apache-2.0 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 5e-05\n* train\\_batch\\_size: 8\n* eval\\_batch\\_size: 8\n* seed: 42\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* num\\_epochs: 3", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* Pytorch 2.2.1+cu121\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]
question-answering	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # my_awesome_qa_model This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 1.5822 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 3 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| \|:-------------:\|:-----:\|:----:\|:---------------:\| \| No log \| 1.0 \| 250 \| 2.3558 \| \| 2.7849 \| 2.0 \| 500 \| 1.6460 \| \| 2.7849 \| 3.0 \| 750 \| 1.5822 \| ### Framework versions - Transformers 4.39.3 - Pytorch 2.2.2+cu118 - Datasets 2.18.0 - Tokenizers 0.15.2	{"license": "apache-2.0", "tags": ["generated_from_trainer"], "base_model": "distilbert-base-uncased", "model-index": [{"name": "my_awesome_qa_model", "results": []}]}	sritharansk/my_awesome_qa_model	null	[ "transformers", "tensorboard", "safetensors", "distilbert", "question-answering", "generated_from_trainer", "base_model:distilbert-base-uncased", "license:apache-2.0", "endpoints_compatible", "region:us" ]	null	2024-04-19T04:06:18+00:00	[]	[]	TAGS #transformers #tensorboard #safetensors #distilbert #question-answering #generated_from_trainer #base_model-distilbert-base-uncased #license-apache-2.0 #endpoints_compatible #region-us	my\_awesome\_qa\_model ====================== This model is a fine-tuned version of distilbert-base-uncased on an unknown dataset. It achieves the following results on the evaluation set: * Loss: 1.5822 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * learning\_rate: 2e-05 * train\_batch\_size: 16 * eval\_batch\_size: 16 * seed: 42 * optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 * lr\_scheduler\_type: linear * num\_epochs: 3 ### Training results ### Framework versions * Transformers 4.39.3 * Pytorch 2.2.2+cu118 * Datasets 2.18.0 * Tokenizers 0.15.2	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 2e-05\n* train\\_batch\\_size: 16\n* eval\\_batch\\_size: 16\n* seed: 42\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* num\\_epochs: 3", "### Training results", "### Framework versions\n\n\n* Transformers 4.39.3\n* Pytorch 2.2.2+cu118\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #tensorboard #safetensors #distilbert #question-answering #generated_from_trainer #base_model-distilbert-base-uncased #license-apache-2.0 #endpoints_compatible #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 2e-05\n* train\\_batch\\_size: 16\n* eval\\_batch\\_size: 16\n* seed: 42\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* num\\_epochs: 3", "### Training results", "### Framework versions\n\n\n* Transformers 4.39.3\n* Pytorch 2.2.2+cu118\n* Datasets 2.18.0\n* Tokenizers 0.15.2" ]
null	peft	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # llama-3-8b-it-fr This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 1 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 4 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 2 - training_steps: 150 - mixed_precision_training: Native AMP ### Framework versions - PEFT 0.8.2 - Transformers 4.38.0 - Pytorch 2.2.1+cu121 - Datasets 2.17.0 - Tokenizers 0.15.2	{"license": "other", "library_name": "peft", "tags": ["trl", "sft", "generated_from_trainer"], "base_model": "meta-llama/Meta-Llama-3-8B-Instruct", "model-index": [{"name": "llama-3-8b-it-fr", "results": []}]}	1rsh/llama-3-8b-it-fr	null	[ "peft", "tensorboard", "safetensors", "trl", "sft", "generated_from_trainer", "base_model:meta-llama/Meta-Llama-3-8B-Instruct", "license:other", "region:us" ]	null	2024-04-19T04:07:00+00:00	[]	[]	TAGS #peft #tensorboard #safetensors #trl #sft #generated_from_trainer #base_model-meta-llama/Meta-Llama-3-8B-Instruct #license-other #region-us	# llama-3-8b-it-fr This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 1 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 4 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 2 - training_steps: 150 - mixed_precision_training: Native AMP ### Framework versions - PEFT 0.8.2 - Transformers 4.38.0 - Pytorch 2.2.1+cu121 - Datasets 2.17.0 - Tokenizers 0.15.2	[ "# llama-3-8b-it-fr\n\nThis model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on an unknown dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 0.0002\n- train_batch_size: 1\n- eval_batch_size: 8\n- seed: 42\n- gradient_accumulation_steps: 4\n- total_train_batch_size: 4\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- lr_scheduler_warmup_steps: 2\n- training_steps: 150\n- mixed_precision_training: Native AMP", "### Framework versions\n\n- PEFT 0.8.2\n- Transformers 4.38.0\n- Pytorch 2.2.1+cu121\n- Datasets 2.17.0\n- Tokenizers 0.15.2" ]	[ "TAGS\n#peft #tensorboard #safetensors #trl #sft #generated_from_trainer #base_model-meta-llama/Meta-Llama-3-8B-Instruct #license-other #region-us \n", "# llama-3-8b-it-fr\n\nThis model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on an unknown dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 0.0002\n- train_batch_size: 1\n- eval_batch_size: 8\n- seed: 42\n- gradient_accumulation_steps: 4\n- total_train_batch_size: 4\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- lr_scheduler_warmup_steps: 2\n- training_steps: 150\n- mixed_precision_training: Native AMP", "### Framework versions\n\n- PEFT 0.8.2\n- Transformers 4.38.0\n- Pytorch 2.2.1+cu121\n- Datasets 2.17.0\n- Tokenizers 0.15.2" ]
image-classification	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # swiftformer-xs-dmae-va-U5-42C This model is a fine-tuned version of [MBZUAI/swiftformer-xs](https://huggingface.co/MBZUAI/swiftformer-xs) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 0.6744 - Accuracy: 0.8167 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 32 - eval_batch_size: 32 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 128 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 80 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\| \| No log \| 0.9 \| 7 \| 1.3856 \| 0.4667 \| \| 1.3855 \| 1.94 \| 15 \| 1.3819 \| 0.4833 \| \| 1.3855 \| 2.97 \| 23 \| 1.3687 \| 0.4333 \| \| 1.3742 \| 4.0 \| 31 \| 1.3189 \| 0.3167 \| \| 1.3004 \| 4.9 \| 38 \| 1.2501 \| 0.4833 \| \| 1.3004 \| 5.94 \| 46 \| 1.2268 \| 0.4833 \| \| 1.1716 \| 6.97 \| 54 \| 1.2115 \| 0.5 \| \| 1.0686 \| 8.0 \| 62 \| 1.2243 \| 0.5333 \| \| 1.0686 \| 8.9 \| 69 \| 1.1432 \| 0.55 \| \| 0.9764 \| 9.94 \| 77 \| 1.0205 \| 0.55 \| \| 0.873 \| 10.97 \| 85 \| 0.9721 \| 0.6 \| \| 0.873 \| 12.0 \| 93 \| 0.9221 \| 0.5667 \| \| 0.7822 \| 12.9 \| 100 \| 0.8593 \| 0.6167 \| \| 0.664 \| 13.94 \| 108 \| 0.7775 \| 0.7 \| \| 0.664 \| 14.97 \| 116 \| 0.8117 \| 0.6167 \| \| 0.5439 \| 16.0 \| 124 \| 0.7553 \| 0.6833 \| \| 0.5439 \| 16.9 \| 131 \| 0.6697 \| 0.7167 \| \| 0.496 \| 17.94 \| 139 \| 0.6480 \| 0.7333 \| \| 0.4563 \| 18.97 \| 147 \| 0.7115 \| 0.7333 \| \| 0.4563 \| 20.0 \| 155 \| 0.6777 \| 0.7167 \| \| 0.3831 \| 20.9 \| 162 \| 0.6416 \| 0.7667 \| \| 0.339 \| 21.94 \| 170 \| 0.7040 \| 0.7 \| \| 0.339 \| 22.97 \| 178 \| 0.6859 \| 0.7167 \| \| 0.3033 \| 24.0 \| 186 \| 0.6012 \| 0.7 \| \| 0.2655 \| 24.9 \| 193 \| 0.5440 \| 0.7833 \| \| 0.2655 \| 25.94 \| 201 \| 0.6174 \| 0.75 \| \| 0.2269 \| 26.97 \| 209 \| 0.5745 \| 0.7333 \| \| 0.2472 \| 28.0 \| 217 \| 0.5688 \| 0.8 \| \| 0.2472 \| 28.9 \| 224 \| 0.6578 \| 0.75 \| \| 0.2004 \| 29.94 \| 232 \| 0.5811 \| 0.7833 \| \| 0.2099 \| 30.97 \| 240 \| 0.6672 \| 0.75 \| \| 0.2099 \| 32.0 \| 248 \| 0.5927 \| 0.75 \| \| 0.1834 \| 32.9 \| 255 \| 0.6193 \| 0.7667 \| \| 0.1834 \| 33.94 \| 263 \| 0.7505 \| 0.7167 \| \| 0.2248 \| 34.97 \| 271 \| 0.7730 \| 0.7333 \| \| 0.1571 \| 36.0 \| 279 \| 0.6211 \| 0.7333 \| \| 0.1571 \| 36.9 \| 286 \| 0.6228 \| 0.7167 \| \| 0.1983 \| 37.94 \| 294 \| 0.6088 \| 0.7167 \| \| 0.1629 \| 38.97 \| 302 \| 0.7009 \| 0.75 \| \| 0.1629 \| 40.0 \| 310 \| 0.7285 \| 0.75 \| \| 0.1547 \| 40.9 \| 317 \| 0.6401 \| 0.7667 \| \| 0.1548 \| 41.94 \| 325 \| 0.6123 \| 0.7833 \| \| 0.1548 \| 42.97 \| 333 \| 0.6317 \| 0.8 \| \| 0.1566 \| 44.0 \| 341 \| 0.7579 \| 0.7167 \| \| 0.1361 \| 44.9 \| 348 \| 0.6653 \| 0.7167 \| \| 0.1361 \| 45.94 \| 356 \| 0.7401 \| 0.7333 \| \| 0.1273 \| 46.97 \| 364 \| 0.8404 \| 0.7333 \| \| 0.1312 \| 48.0 \| 372 \| 0.8388 \| 0.75 \| \| 0.1312 \| 48.9 \| 379 \| 0.7823 \| 0.7667 \| \| 0.1307 \| 49.94 \| 387 \| 0.6980 \| 0.7167 \| \| 0.1307 \| 50.97 \| 395 \| 0.7589 \| 0.75 \| \| 0.1061 \| 52.0 \| 403 \| 0.6644 \| 0.7333 \| \| 0.1186 \| 52.9 \| 410 \| 0.7057 \| 0.7333 \| \| 0.1186 \| 53.94 \| 418 \| 0.6744 \| 0.8167 \| \| 0.1108 \| 54.97 \| 426 \| 0.6328 \| 0.7667 \| \| 0.1014 \| 56.0 \| 434 \| 0.6402 \| 0.7833 \| \| 0.1014 \| 56.9 \| 441 \| 0.6631 \| 0.75 \| \| 0.1082 \| 57.94 \| 449 \| 0.7001 \| 0.7333 \| \| 0.1118 \| 58.97 \| 457 \| 0.7898 \| 0.7333 \| \| 0.1118 \| 60.0 \| 465 \| 0.7644 \| 0.7333 \| \| 0.1051 \| 60.9 \| 472 \| 0.7767 \| 0.7333 \| \| 0.0979 \| 61.94 \| 480 \| 0.7440 \| 0.7333 \| \| 0.0979 \| 62.97 \| 488 \| 0.6827 \| 0.7333 \| \| 0.0834 \| 64.0 \| 496 \| 0.7008 \| 0.7333 \| \| 0.0834 \| 64.9 \| 503 \| 0.7243 \| 0.7167 \| \| 0.0963 \| 65.94 \| 511 \| 0.7656 \| 0.7333 \| \| 0.0989 \| 66.97 \| 519 \| 0.7332 \| 0.7333 \| \| 0.0989 \| 68.0 \| 527 \| 0.7624 \| 0.7333 \| \| 0.107 \| 68.9 \| 534 \| 0.7292 \| 0.75 \| \| 0.0987 \| 69.94 \| 542 \| 0.7169 \| 0.7333 \| \| 0.0987 \| 70.97 \| 550 \| 0.7462 \| 0.7333 \| \| 0.0956 \| 72.0 \| 558 \| 0.6656 \| 0.75 \| \| 0.0956 \| 72.26 \| 560 \| 0.6873 \| 0.7333 \| ### Framework versions - Transformers 4.36.2 - Pytorch 2.1.2+cu118 - Datasets 2.16.1 - Tokenizers 0.15.0	{"tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "MBZUAI/swiftformer-xs", "model-index": [{"name": "swiftformer-xs-dmae-va-U5-42C", "results": []}]}	Augusto777/swiftformer-xs-dmae-va-U5-80C	null	[ "transformers", "tensorboard", "safetensors", "swiftformer", "image-classification", "generated_from_trainer", "base_model:MBZUAI/swiftformer-xs", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T04:08:26+00:00	[]	[]	TAGS #transformers #tensorboard #safetensors #swiftformer #image-classification #generated_from_trainer #base_model-MBZUAI/swiftformer-xs #autotrain_compatible #endpoints_compatible #region-us	swiftformer-xs-dmae-va-U5-42C ============================= This model is a fine-tuned version of MBZUAI/swiftformer-xs on an unknown dataset. It achieves the following results on the evaluation set: * Loss: 0.6744 * Accuracy: 0.8167 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * learning\_rate: 0.0002 * train\_batch\_size: 32 * eval\_batch\_size: 32 * seed: 42 * gradient\_accumulation\_steps: 4 * total\_train\_batch\_size: 128 * optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 * lr\_scheduler\_type: linear * lr\_scheduler\_warmup\_ratio: 0.1 * num\_epochs: 80 ### Training results ### Framework versions * Transformers 4.36.2 * Pytorch 2.1.2+cu118 * Datasets 2.16.1 * Tokenizers 0.15.0	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 0.0002\n* train\\_batch\\_size: 32\n* eval\\_batch\\_size: 32\n* seed: 42\n* gradient\\_accumulation\\_steps: 4\n* total\\_train\\_batch\\_size: 128\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* lr\\_scheduler\\_warmup\\_ratio: 0.1\n* num\\_epochs: 80", "### Training results", "### Framework versions\n\n\n* Transformers 4.36.2\n* Pytorch 2.1.2+cu118\n* Datasets 2.16.1\n* Tokenizers 0.15.0" ]	[ "TAGS\n#transformers #tensorboard #safetensors #swiftformer #image-classification #generated_from_trainer #base_model-MBZUAI/swiftformer-xs #autotrain_compatible #endpoints_compatible #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 0.0002\n* train\\_batch\\_size: 32\n* eval\\_batch\\_size: 32\n* seed: 42\n* gradient\\_accumulation\\_steps: 4\n* total\\_train\\_batch\\_size: 128\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* lr\\_scheduler\\_warmup\\_ratio: 0.1\n* num\\_epochs: 80", "### Training results", "### Framework versions\n\n\n* Transformers 4.36.2\n* Pytorch 2.1.2+cu118\n* Datasets 2.16.1\n* Tokenizers 0.15.0" ]
text-generation	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	heyllm234/sc37	null	[ "transformers", "safetensors", "stablelm", "text-generation", "conversational", "arxiv:1910.09700", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T04:08:30+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #stablelm #text-generation #conversational #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #stablelm #text-generation #conversational #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
text-generation	transformers	<!-- header start --> <!-- 200823 --> <div style="width: auto; margin-left: auto; margin-right: auto"> <img src="https://www.astronomer.io/logo/astronomer-logo-RGB-standard-1200px.png" alt="Astronomer" style="width: 60%; min-width: 400px; display: block; margin: auto;"> </div> <div style="margin-top: 1.0em; margin-bottom: 1.0em;"></div> <div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">This model is generously created and made open source by <a href="https://astronomer.io">Astronomer</a>.</p></div> <div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">Astronomer is the de facto company for <a href="https://airflow.apache.org/">Apache Airflow</a>, the most trusted open-source framework for data orchestration and MLOps.</p></div> <hr style="margin-top: 1.0em; margin-bottom: 1.0em;"> <!-- header end --> # Llama-3-8B-Instruct-GPTQ-8-Bit - Original Model creator: [Meta Llama from Meta](https://huggingface.co/meta-llama) - Original model: [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) - Built with Meta Llama 3 - Quantized by [Astronomer](https://astronomer.io) # Important Note About Serving with vLLM & oobabooga/text-generation-webui - For loading this model onto vLLM, make sure all requests have `"stop_token_ids":[128001, 128009]` to temporarily address the non-stop generation issue. - vLLM does not yet respect `generation_config.json`. - vLLM team is working on a a fix for this https://github.com/vllm-project/vllm/issues/4180 - For oobabooga/text-generation-webui - Load the model via AutoGPTQ, with `no_inject_fused_attention` enabled. This is a bug with AutoGPTQ library. - Under `Parameters` -> `Generation` -> `Skip special tokens`: turn this off (deselect) - Under `Parameters` -> `Generation` -> `Custom stopping strings`: add `"<\|end_of_text\|>","<\|eot_id\|>"` to the field <!-- description start --> ## Description This repo contains 8 Bit quantized GPTQ model files for [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct). This model can be loaded with just over 10GB of VRAM (compared to the original 16.07GB model) and can be served lightning fast with the cheapest Nvidia GPUs possible (Nvidia T4, Nvidia K80, RTX 4070, etc). The 8 bit GPTQ quant has minimum quality degradation from the original `bfloat16` model due to its higher bitrate. <!-- description end --> ## GPTQ Quantization Method - This model is quantized by utilizing the AutoGPTQ library, following best practices noted by [GPTQ paper](https://arxiv.org/abs/2210.17323) - Quantization is calibrated and aligned with random samples from the specified dataset (wikitext for now) for minimum accuracy loss. \| Branch \| Bits \| Group Size \| Act Order \| Damp % \| GPTQ Dataset \| Sequence Length \| VRAM Size \| ExLlama \| Description \| \| ------ \| ---- \| -- \| --------- \| ------ \| ------------ \| ------- \| ---- \| ------- \| ---- \| \| [main](https://huggingface.co/astronomer-io/Llama-3-8B-Instruct-GPTQ-8-Bit/tree/main) \| 8 \| 32 \| Yes \| 0.1 \| [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) \| 8192 \| 9.09 GB \| No \| 8-bit, with Act Order and group size 32g. Minimum accuracy loss with decent VRAM usage reduction. \| \| More variants to come \| TBD \| TBD \| TBD \| TBD \| TBD \| TBD \| TBD \| TBD \| May upload additional variants of GPTQ 8 bit models in the future using different parameters such as 128g group size and etc. \| ## Serving this GPTQ model using vLLM Tested serving this model via vLLM using an Nvidia T4 (16GB VRAM). Tested with the below command ``` python -m vllm.entrypoints.openai.api_server --model astronomer-io/Llama-3-8B-Instruct-GPTQ-8-Bit --max-model-len 8192 --dtype float16 ``` For the non-stop token generation bug, make sure to send requests with `stop_token_ids":[128001, 128009]` to vLLM endpoint Example: ```json { "model": "astronomer-io/Llama-3-8B-Instruct-GPTQ-8-Bit", "messages": [ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": "Who created Llama 3?"} ], "max_tokens": 2000, "stop_token_ids":[128001,128009] } ``` ### Prompt Template ``` <\|begin_of_text\|><\|start_header_id\|>user<\|end_header_id\|> {{prompt}}<\|eot_id\|> <\|start_header_id\|>assistant<\|end_header_id\|> ``` ### Contributors - Quantized by [David Xue, Machine Learning Engineer from Astronomer](https://www.linkedin.com/in/david-xue-uva/)	{"license": "other", "tags": ["llama", "llama-3", "facebook", "meta", "astronomer", "gptq", "pretrained", "quantized", "finetuned", "autotrain_compatible", "endpoints_compatible"], "datasets": ["wikitext"], "model_name": "Meta-Llama-3-8B-Instruct", "base_model": "meta-llama/Meta-Llama-3-8B-Instruct", "inference": false, "model_creator": "astronomer-io", "model_type": "llama", "pipeline_tag": "text-generation", "prompt_template": "{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<\|start_header_id\|>' + message['role'] + '<\|end_header_id\|>\n\n'+ message['content'] \| trim + '<\|eot_id\|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<\|start_header_id\|>assistant<\|end_header_id\|>\n\n' }}{% endif %}", "quantized_by": "davidxmle", "license_name": "llama-3-community-license", "license_link": "https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct/blob/main/LICENSE"}	astronomer/Llama-3-8B-Instruct-GPTQ-8-Bit	null	[ "transformers", "safetensors", "llama", "text-generation", "llama-3", "facebook", "meta", "astronomer", "gptq", "pretrained", "quantized", "finetuned", "autotrain_compatible", "endpoints_compatible", "conversational", "dataset:wikitext", "arxiv:2210.17323", "base_model:meta-llama/Meta-Llama-3-8B-Instruct", "license:other", "text-generation-inference", "8-bit", "region:us" ]	null	2024-04-19T04:08:47+00:00	[ "2210.17323" ]	[]	TAGS #transformers #safetensors #llama #text-generation #llama-3 #facebook #meta #astronomer #gptq #pretrained #quantized #finetuned #autotrain_compatible #endpoints_compatible #conversational #dataset-wikitext #arxiv-2210.17323 #base_model-meta-llama/Meta-Llama-3-8B-Instruct #license-other #text-generation-inference #8-bit #region-us	![](URL alt=) This model is generously created and made open source by [Astronomer](URL). Astronomer is the de facto company for [Llama-3-8B-Instruct-GPTQ-8-Bit ============================== * Original Model creator: Meta Llama from Meta * Original model: meta-llama/Meta-Llama-3-8B-Instruct * Built with Meta Llama 3 * Quantized by Astronomer Important Note About Serving with vLLM & oobabooga/text-generation-webui ======================================================================== * For loading this model onto vLLM, make sure all requests have '"stop\_token\_ids":[128001, 128009]' to temporarily address the non-stop generation issue. + vLLM does not yet respect 'generation\_config.json'. + vLLM team is working on a a fix for this URL * For oobabooga/text-generation-webui + Load the model via AutoGPTQ, with 'no\_inject\_fused\_attention' enabled. This is a bug with AutoGPTQ library. + Under 'Parameters' -> 'Generation' -> 'Skip special tokens': turn this off (deselect) + Under 'Parameters' -> 'Generation' -> 'Custom stopping strings': add '"<\|end\_of\_text\|>","<\|eot\_id\|>"' to the field Description ----------- This repo contains 8 Bit quantized GPTQ model files for meta-llama/Meta-Llama-3-8B-Instruct. This model can be loaded with just over 10GB of VRAM (compared to the original 16.07GB model) and can be served lightning fast with the cheapest Nvidia GPUs possible (Nvidia T4, Nvidia K80, RTX 4070, etc). The 8 bit GPTQ quant has minimum quality degradation from the original 'bfloat16' model due to its higher bitrate. GPTQ Quantization Method ------------------------ * This model is quantized by utilizing the AutoGPTQ library, following best practices noted by GPTQ paper * Quantization is calibrated and aligned with random samples from the specified dataset (wikitext for now) for minimum accuracy loss. Serving this GPTQ model using vLLM ---------------------------------- Tested serving this model via vLLM using an Nvidia T4 (16GB VRAM). Tested with the below command For the non-stop token generation bug, make sure to send requests with 'stop\_token\_ids":[128001, 128009]' to vLLM endpoint Example: ### Prompt Template ### Contributors * Quantized by David Xue, Machine Learning Engineer from Astronomer](URL Airflow</a>, the most trusted open-source framework for data orchestration and MLOps.</p></div> <hr style=)	[ "### Prompt Template", "### Contributors\n\n\n* Quantized by David Xue, Machine Learning Engineer from Astronomer](URL Airflow</a>, the most trusted open-source framework for data orchestration and MLOps.</p></div>\n<hr style=)" ]	[ "TAGS\n#transformers #safetensors #llama #text-generation #llama-3 #facebook #meta #astronomer #gptq #pretrained #quantized #finetuned #autotrain_compatible #endpoints_compatible #conversational #dataset-wikitext #arxiv-2210.17323 #base_model-meta-llama/Meta-Llama-3-8B-Instruct #license-other #text-generation-inference #8-bit #region-us \n", "### Prompt Template", "### Contributors\n\n\n* Quantized by David Xue, Machine Learning Engineer from Astronomer](URL Airflow</a>, the most trusted open-source framework for data orchestration and MLOps.</p></div>\n<hr style=)" ]
audio-classification	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # hubert_classifier_arabic This model is a fine-tuned version of [facebook/hubert-base-ls960](https://huggingface.co/facebook/hubert-base-ls960) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 2.2809 - Accuracy: 0.3852 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 3e-05 - train_batch_size: 32 - eval_batch_size: 32 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 128 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.01 - training_steps: 2000 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\| \| 4.214 \| 1.0 \| 219 \| 3.6941 \| 0.0767 \| \| 3.6773 \| 2.0 \| 438 \| 3.3037 \| 0.0921 \| \| 3.4196 \| 3.0 \| 657 \| 3.0830 \| 0.1457 \| \| 3.2171 \| 4.0 \| 876 \| 2.8489 \| 0.2231 \| \| 3.0634 \| 5.0 \| 1095 \| 2.6796 \| 0.2812 \| \| 2.8922 \| 6.0 \| 1314 \| 2.5014 \| 0.3226 \| \| 2.7706 \| 7.0 \| 1533 \| 2.4051 \| 0.3409 \| \| 2.6681 \| 8.0 \| 1752 \| 2.2995 \| 0.3811 \| \| 2.6042 \| 9.0 \| 1971 \| 2.2819 \| 0.3820 \| \| 2.562 \| 9.13 \| 2000 \| 2.2809 \| 0.3852 \| ### Framework versions - Transformers 4.33.3 - Pytorch 2.2.1 - Datasets 2.18.0 - Tokenizers 0.13.3	{"license": "apache-2.0", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "facebook/hubert-base-ls960", "model-index": [{"name": "hubert_classifier_arabic", "results": []}]}	fydhfzh/hubert_classifier_arabic	null	[ "transformers", "pytorch", "hubert", "audio-classification", "generated_from_trainer", "base_model:facebook/hubert-base-ls960", "license:apache-2.0", "endpoints_compatible", "region:us" ]	null	2024-04-19T04:14:14+00:00	[]	[]	TAGS #transformers #pytorch #hubert #audio-classification #generated_from_trainer #base_model-facebook/hubert-base-ls960 #license-apache-2.0 #endpoints_compatible #region-us	hubert\_classifier\_arabic ========================== This model is a fine-tuned version of facebook/hubert-base-ls960 on an unknown dataset. It achieves the following results on the evaluation set: * Loss: 2.2809 * Accuracy: 0.3852 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * learning\_rate: 3e-05 * train\_batch\_size: 32 * eval\_batch\_size: 32 * seed: 42 * gradient\_accumulation\_steps: 4 * total\_train\_batch\_size: 128 * optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 * lr\_scheduler\_type: linear * lr\_scheduler\_warmup\_ratio: 0.01 * training\_steps: 2000 ### Training results ### Framework versions * Transformers 4.33.3 * Pytorch 2.2.1 * Datasets 2.18.0 * Tokenizers 0.13.3	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 3e-05\n* train\\_batch\\_size: 32\n* eval\\_batch\\_size: 32\n* seed: 42\n* gradient\\_accumulation\\_steps: 4\n* total\\_train\\_batch\\_size: 128\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* lr\\_scheduler\\_warmup\\_ratio: 0.01\n* training\\_steps: 2000", "### Training results", "### Framework versions\n\n\n* Transformers 4.33.3\n* Pytorch 2.2.1\n* Datasets 2.18.0\n* Tokenizers 0.13.3" ]	[ "TAGS\n#transformers #pytorch #hubert #audio-classification #generated_from_trainer #base_model-facebook/hubert-base-ls960 #license-apache-2.0 #endpoints_compatible #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 3e-05\n* train\\_batch\\_size: 32\n* eval\\_batch\\_size: 32\n* seed: 42\n* gradient\\_accumulation\\_steps: 4\n* total\\_train\\_batch\\_size: 128\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* lr\\_scheduler\\_warmup\\_ratio: 0.01\n* training\\_steps: 2000", "### Training results", "### Framework versions\n\n\n* Transformers 4.33.3\n* Pytorch 2.2.1\n* Datasets 2.18.0\n* Tokenizers 0.13.3" ]
question-answering	peft	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # Mistral-Alpaca-Finetuned This model is a fine-tuned version of [TheBloke/Mistral-7B-Instruct-v0.1-GPTQ](https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.1-GPTQ) on the tatsu-lab/alpaca dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - num_epochs: 1 - mixed_precision_training: Native AMP ### Framework versions - PEFT 0.10.0 - Transformers 4.41.0.dev0 - Pytorch 2.2.1+cu121 - Datasets 2.18.0 - Tokenizers 0.19.1	{"license": "apache-2.0", "library_name": "peft", "tags": ["trl", "sft", "generated_from_trainer"], "datasets": ["tatsu-lab/alpaca"], "base_model": "TheBloke/Mistral-7B-Instruct-v0.1-GPTQ", "pipeline_tag": "question-answering", "model-index": [{"name": "Mistral-Alpaca-Finetuned", "results": []}]}	karthiksagarn/Mistral-Alpaca-Finetuned	null	[ "peft", "tensorboard", "safetensors", "trl", "sft", "generated_from_trainer", "question-answering", "dataset:tatsu-lab/alpaca", "base_model:TheBloke/Mistral-7B-Instruct-v0.1-GPTQ", "license:apache-2.0", "region:us" ]	null	2024-04-19T04:17:31+00:00	[]	[]	TAGS #peft #tensorboard #safetensors #trl #sft #generated_from_trainer #question-answering #dataset-tatsu-lab/alpaca #base_model-TheBloke/Mistral-7B-Instruct-v0.1-GPTQ #license-apache-2.0 #region-us	# Mistral-Alpaca-Finetuned This model is a fine-tuned version of TheBloke/Mistral-7B-Instruct-v0.1-GPTQ on the tatsu-lab/alpaca dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - num_epochs: 1 - mixed_precision_training: Native AMP ### Framework versions - PEFT 0.10.0 - Transformers 4.41.0.dev0 - Pytorch 2.2.1+cu121 - Datasets 2.18.0 - Tokenizers 0.19.1	[ "# Mistral-Alpaca-Finetuned\n\nThis model is a fine-tuned version of TheBloke/Mistral-7B-Instruct-v0.1-GPTQ on the tatsu-lab/alpaca dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 0.0002\n- train_batch_size: 8\n- eval_batch_size: 8\n- seed: 42\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: cosine\n- num_epochs: 1\n- mixed_precision_training: Native AMP", "### Framework versions\n\n- PEFT 0.10.0\n- Transformers 4.41.0.dev0\n- Pytorch 2.2.1+cu121\n- Datasets 2.18.0\n- Tokenizers 0.19.1" ]	[ "TAGS\n#peft #tensorboard #safetensors #trl #sft #generated_from_trainer #question-answering #dataset-tatsu-lab/alpaca #base_model-TheBloke/Mistral-7B-Instruct-v0.1-GPTQ #license-apache-2.0 #region-us \n", "# Mistral-Alpaca-Finetuned\n\nThis model is a fine-tuned version of TheBloke/Mistral-7B-Instruct-v0.1-GPTQ on the tatsu-lab/alpaca dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 0.0002\n- train_batch_size: 8\n- eval_batch_size: 8\n- seed: 42\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: cosine\n- num_epochs: 1\n- mixed_precision_training: Native AMP", "### Framework versions\n\n- PEFT 0.10.0\n- Transformers 4.41.0.dev0\n- Pytorch 2.2.1+cu121\n- Datasets 2.18.0\n- Tokenizers 0.19.1" ]
null	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	RajSang/hbert	null	[ "transformers", "arxiv:1910.09700", "endpoints_compatible", "region:us" ]	null	2024-04-19T04:17:59+00:00	[ "1910.09700" ]	[]	TAGS #transformers #arxiv-1910.09700 #endpoints_compatible #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #arxiv-1910.09700 #endpoints_compatible #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
text2text-generation	null	Here is an example of how to use this model: https://github.com/LJ-Hao/MLC-LLM-on-Jetson >Note: This model only support Jetson Containers with MLC: https://github.com/dusty-nv/jetson-containers ``` from mlc_chat import ChatModule from mlc_chat.callback import StreamToStdout def run(): cm = ChatModule( model="PATH/params", model_lib_path="PATH/Llama-2-7b-chat-hf-q4f16_ft-cuda.so") while True: try: prompt = input('\033[94m' +"Prompt: " + '\033[0m') cm.generate( prompt=prompt, progress_callback=StreamToStdout(callback_interval=2), ) except KeyboardInterrupt: break if __name__ == '__main__': run() ```	{"license": "mit", "pipeline_tag": "text2text-generation"}	JiahaoLi/llama2-7b-MLC-q4f16-jetson-containers	null	[ "text2text-generation", "license:mit", "region:us" ]	null	2024-04-19T04:18:16+00:00	[]	[]	TAGS #text2text-generation #license-mit #region-us	Here is an example of how to use this model: URL >Note: This model only support Jetson Containers with MLC: URL	[]	[ "TAGS\n#text2text-generation #license-mit #region-us \n" ]
null	peft	## Training procedure The following `bitsandbytes` quantization config was used during training: - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: True - bnb_4bit_compute_dtype: bfloat16 ### Framework versions - PEFT 0.4.0	{"library_name": "peft"}	nuggi/sql-classification-llama-2-7b-test	null	[ "peft", "pytorch", "llama", "region:us" ]	null	2024-04-19T04:18:55+00:00	[]	[]	TAGS #peft #pytorch #llama #region-us	## Training procedure The following 'bitsandbytes' quantization config was used during training: - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: True - bnb_4bit_compute_dtype: bfloat16 ### Framework versions - PEFT 0.4.0	[ "## Training procedure\n\n\nThe following 'bitsandbytes' quantization config was used during training:\n- load_in_8bit: False\n- load_in_4bit: True\n- llm_int8_threshold: 6.0\n- llm_int8_skip_modules: None\n- llm_int8_enable_fp32_cpu_offload: False\n- llm_int8_has_fp16_weight: False\n- bnb_4bit_quant_type: nf4\n- bnb_4bit_use_double_quant: True\n- bnb_4bit_compute_dtype: bfloat16", "### Framework versions\n\n\n- PEFT 0.4.0" ]	[ "TAGS\n#peft #pytorch #llama #region-us \n", "## Training procedure\n\n\nThe following 'bitsandbytes' quantization config was used during training:\n- load_in_8bit: False\n- load_in_4bit: True\n- llm_int8_threshold: 6.0\n- llm_int8_skip_modules: None\n- llm_int8_enable_fp32_cpu_offload: False\n- llm_int8_has_fp16_weight: False\n- bnb_4bit_quant_type: nf4\n- bnb_4bit_use_double_quant: True\n- bnb_4bit_compute_dtype: bfloat16", "### Framework versions\n\n\n- PEFT 0.4.0" ]
text-generation	transformers	This is the first version of upscaling llama-3. Version 2 is now out and does not have any of the issues that this version has. Please use version 2 instead. Linked bellow: - https://huggingface.co/Replete-AI/Llama-3-11.5B-v2 __________________________________________________________________ Llama-3-13B Thank you to Meta for the weights for Meta-Llama-3-8B ![image/png](https://cdn-uploads.huggingface.co/production/uploads/642cc1c253e76b4c2286c58e/aJJxKus1wP5N-euvHEUq7.png) This is an upscaling of the Llama-3-8B Ai using techniques created for Mistral-Evolved-11b-v0.1. This Ai model has been upscaled from 8b parameters to 13b parameters without any continuous pretraining or fine-tuning. From testing, the model seems to function perfectly at fp16, but has some issues at 4-bit quantization using bitsandbytes. The model that was used to create this one is linked below: https://huggingface.co/meta-llama/Meta-Llama-3-8B - Llama-3-13B \| Metric \|Value\| \|---------------------------------\|----:\| \|Avg. \|54.61\| \|AI2 Reasoning Challenge (25-Shot)\|52.99\| \|HellaSwag (10-Shot) \|80.66\| \|MMLU (5-Shot) \|62.12\| \|TruthfulQA (0-shot) \|39.28\| \|Winogrande (5-shot) \|70.72\| \|GSM8k (5-shot) \|21.91\| - Original Meta-Llama-3-8B \| Metric \|Value\| \|---------------------------------\|----:\| \|Avg. \|62.87\| \|AI2 Reasoning Challenge (25-Shot)\|59.47\| \|HellaSwag (10-Shot) \|82.09\| \|MMLU (5-Shot) \|66.69\| \|TruthfulQA (0-shot) \|43.90\| \|Winogrande (5-shot) \|77.35\| \|GSM8k (5-shot) \|45.34\|	{"license": "other", "library_name": "transformers", "base_model": [], "license_name": "llama-3", "license_link": "https://llama.meta.com/llama3/license/", "thumbnail": "https://cdn-uploads.huggingface.co/production/uploads/642cc1c253e76b4c2286c58e/aJJxKus1wP5N-euvHEUq7.png"}	Replete-AI/Llama-3-13B	null	[ "transformers", "safetensors", "llama", "text-generation", "conversational", "license:other", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T04:20:05+00:00	[]	[]	TAGS #transformers #safetensors #llama #text-generation #conversational #license-other #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	This is the first version of upscaling llama-3. Version 2 is now out and does not have any of the issues that this version has. Please use version 2 instead. Linked bellow: * URL --- Llama-3-13B Thank you to Meta for the weights for Meta-Llama-3-8B !image/png This is an upscaling of the Llama-3-8B Ai using techniques created for Mistral-Evolved-11b-v0.1. This Ai model has been upscaled from 8b parameters to 13b parameters without any continuous pretraining or fine-tuning. From testing, the model seems to function perfectly at fp16, but has some issues at 4-bit quantization using bitsandbytes. The model that was used to create this one is linked below: URL * Llama-3-13B * Original Meta-Llama-3-8B	[]	[ "TAGS\n#transformers #safetensors #llama #text-generation #conversational #license-other #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n" ]
text-classification	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # robust_llm_pythia-31m_ian-022_PasswordMatch_n-its-25 This model is a fine-tuned version of [EleutherAI/pythia-31m](https://huggingface.co/EleutherAI/pythia-31m) on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 8 - eval_batch_size: 64 - seed: 0 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results ### Framework versions - Transformers 4.39.3 - Pytorch 2.2.1 - Datasets 2.18.0 - Tokenizers 0.15.2	{"tags": ["generated_from_trainer"], "base_model": "EleutherAI/pythia-31m", "model-index": [{"name": "robust_llm_pythia-31m_ian-022_PasswordMatch_n-its-25", "results": []}]}	AlignmentResearch/robust_llm_pythia-31m_ian-022_PasswordMatch_n-its-25	null	[ "transformers", "tensorboard", "safetensors", "gpt_neox", "text-classification", "generated_from_trainer", "base_model:EleutherAI/pythia-31m", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T04:20:26+00:00	[]	[]	TAGS #transformers #tensorboard #safetensors #gpt_neox #text-classification #generated_from_trainer #base_model-EleutherAI/pythia-31m #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	# robust_llm_pythia-31m_ian-022_PasswordMatch_n-its-25 This model is a fine-tuned version of EleutherAI/pythia-31m on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 8 - eval_batch_size: 64 - seed: 0 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results ### Framework versions - Transformers 4.39.3 - Pytorch 2.2.1 - Datasets 2.18.0 - Tokenizers 0.15.2	[ "# robust_llm_pythia-31m_ian-022_PasswordMatch_n-its-25\n\nThis model is a fine-tuned version of EleutherAI/pythia-31m on an unknown dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 1e-05\n- train_batch_size: 8\n- eval_batch_size: 64\n- seed: 0\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- num_epochs: 1", "### Training results", "### Framework versions\n\n- Transformers 4.39.3\n- Pytorch 2.2.1\n- Datasets 2.18.0\n- Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #tensorboard #safetensors #gpt_neox #text-classification #generated_from_trainer #base_model-EleutherAI/pythia-31m #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "# robust_llm_pythia-31m_ian-022_PasswordMatch_n-its-25\n\nThis model is a fine-tuned version of EleutherAI/pythia-31m on an unknown dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 1e-05\n- train_batch_size: 8\n- eval_batch_size: 64\n- seed: 0\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- num_epochs: 1", "### Training results", "### Framework versions\n\n- Transformers 4.39.3\n- Pytorch 2.2.1\n- Datasets 2.18.0\n- Tokenizers 0.15.2" ]
null	transformers	# Clevyby/NSM-alpha-7B-Q5_K_S-GGUF This model was converted to GGUF format from [`Nitral-AI/NSM-alpha-7B`](https://huggingface.co/Nitral-AI/NSM-alpha-7B) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space. Refer to the [original model card](https://huggingface.co/Nitral-AI/NSM-alpha-7B) for more details on the model.	{"library_name": "transformers", "tags": ["mergekit", "merge", "llama-cpp", "gguf-my-repo"], "base_model": ["Nitral-AI/Eris_PrimeV3.05-Vision-7B", "Nitral-AI/Nyanade_Stunna-Maid-7B"]}	Clevyby/NSM-alpha-7B-Q5_K_S-GGUF	null	[ "transformers", "gguf", "mergekit", "merge", "llama-cpp", "gguf-my-repo", "base_model:Nitral-AI/Eris_PrimeV3.05-Vision-7B", "base_model:Nitral-AI/Nyanade_Stunna-Maid-7B", "endpoints_compatible", "region:us" ]	null	2024-04-19T04:22:03+00:00	[]	[]	TAGS #transformers #gguf #mergekit #merge #llama-cpp #gguf-my-repo #base_model-Nitral-AI/Eris_PrimeV3.05-Vision-7B #base_model-Nitral-AI/Nyanade_Stunna-Maid-7B #endpoints_compatible #region-us	# Clevyby/NSM-alpha-7B-Q5_K_S-GGUF This model was converted to GGUF format from 'Nitral-AI/NSM-alpha-7B' using URL via the URL's GGUF-my-repo space. Refer to the original model card for more details on the model.	[ "# Clevyby/NSM-alpha-7B-Q5_K_S-GGUF\nThis model was converted to GGUF format from 'Nitral-AI/NSM-alpha-7B' using URL via the URL's GGUF-my-repo space.\nRefer to the original model card for more details on the model." ]	[ "TAGS\n#transformers #gguf #mergekit #merge #llama-cpp #gguf-my-repo #base_model-Nitral-AI/Eris_PrimeV3.05-Vision-7B #base_model-Nitral-AI/Nyanade_Stunna-Maid-7B #endpoints_compatible #region-us \n", "# Clevyby/NSM-alpha-7B-Q5_K_S-GGUF\nThis model was converted to GGUF format from 'Nitral-AI/NSM-alpha-7B' using URL via the URL's GGUF-my-repo space.\nRefer to the original model card for more details on the model." ]
image-classification	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # swinv2-tiny-patch4-window8-256-dmae-va-U5-42B This model is a fine-tuned version of [microsoft/swinv2-tiny-patch4-window8-256](https://huggingface.co/microsoft/swinv2-tiny-patch4-window8-256) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 0.9637 - Accuracy: 0.6667 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 4e-05 - train_batch_size: 32 - eval_batch_size: 32 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 128 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 42 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\| \| No log \| 0.9 \| 7 \| 7.8663 \| 0.1167 \| \| 6.936 \| 1.94 \| 15 \| 7.7572 \| 0.1167 \| \| 6.936 \| 2.97 \| 23 \| 7.1790 \| 0.1167 \| \| 6.7016 \| 4.0 \| 31 \| 5.9033 \| 0.1167 \| \| 5.5439 \| 4.9 \| 38 \| 4.6116 \| 0.1167 \| \| 5.5439 \| 5.94 \| 46 \| 3.2830 \| 0.1167 \| \| 3.6477 \| 6.97 \| 54 \| 2.2014 \| 0.1167 \| \| 2.2506 \| 8.0 \| 62 \| 1.5647 \| 0.45 \| \| 2.2506 \| 8.9 \| 69 \| 1.3160 \| 0.45 \| \| 1.5088 \| 9.94 \| 77 \| 1.3676 \| 0.3333 \| \| 1.3868 \| 10.97 \| 85 \| 1.3390 \| 0.45 \| \| 1.3868 \| 12.0 \| 93 \| 1.3223 \| 0.3833 \| \| 1.351 \| 12.9 \| 100 \| 1.3156 \| 0.45 \| \| 1.3271 \| 13.94 \| 108 \| 1.3485 \| 0.4833 \| \| 1.3271 \| 14.97 \| 116 \| 1.2646 \| 0.4833 \| \| 1.2322 \| 16.0 \| 124 \| 1.2308 \| 0.4833 \| \| 1.2322 \| 16.9 \| 131 \| 1.2160 \| 0.5 \| \| 1.22 \| 17.94 \| 139 \| 1.2015 \| 0.5 \| \| 1.1899 \| 18.97 \| 147 \| 1.2008 \| 0.5 \| \| 1.1899 \| 20.0 \| 155 \| 1.1606 \| 0.5 \| \| 1.109 \| 20.9 \| 162 \| 1.1182 \| 0.5667 \| \| 1.0603 \| 21.94 \| 170 \| 1.0855 \| 0.5333 \| \| 1.0603 \| 22.97 \| 178 \| 1.0763 \| 0.5667 \| \| 1.0264 \| 24.0 \| 186 \| 1.1153 \| 0.5833 \| \| 1.0086 \| 24.9 \| 193 \| 1.0770 \| 0.65 \| \| 1.0086 \| 25.94 \| 201 \| 1.0041 \| 0.6167 \| \| 0.9301 \| 26.97 \| 209 \| 0.9637 \| 0.6667 \| \| 0.9077 \| 28.0 \| 217 \| 0.9824 \| 0.5833 \| \| 0.9077 \| 28.9 \| 224 \| 0.9485 \| 0.6 \| \| 0.8725 \| 29.94 \| 232 \| 0.9294 \| 0.6167 \| \| 0.8203 \| 30.97 \| 240 \| 0.9348 \| 0.6167 \| \| 0.8203 \| 32.0 \| 248 \| 0.9295 \| 0.6 \| \| 0.8211 \| 32.9 \| 255 \| 0.9167 \| 0.6 \| \| 0.8211 \| 33.94 \| 263 \| 0.9281 \| 0.5833 \| \| 0.7916 \| 34.97 \| 271 \| 0.8803 \| 0.6333 \| \| 0.7822 \| 36.0 \| 279 \| 0.8785 \| 0.6333 \| \| 0.7822 \| 36.9 \| 286 \| 0.8906 \| 0.6 \| \| 0.7937 \| 37.94 \| 294 \| 0.8899 \| 0.6 \| ### Framework versions - Transformers 4.36.2 - Pytorch 2.1.2+cu118 - Datasets 2.16.1 - Tokenizers 0.15.0	{"license": "apache-2.0", "tags": ["generated_from_trainer"], "metrics": ["accuracy"], "base_model": "microsoft/swinv2-tiny-patch4-window8-256", "model-index": [{"name": "swinv2-tiny-patch4-window8-256-dmae-va-U5-42B", "results": []}]}	Augusto777/swinv2-tiny-patch4-window8-256-dmae-va-U5-42B	null	[ "transformers", "tensorboard", "safetensors", "swinv2", "image-classification", "generated_from_trainer", "base_model:microsoft/swinv2-tiny-patch4-window8-256", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T04:23:12+00:00	[]	[]	TAGS #transformers #tensorboard #safetensors #swinv2 #image-classification #generated_from_trainer #base_model-microsoft/swinv2-tiny-patch4-window8-256 #license-apache-2.0 #autotrain_compatible #endpoints_compatible #region-us	swinv2-tiny-patch4-window8-256-dmae-va-U5-42B ============================================= This model is a fine-tuned version of microsoft/swinv2-tiny-patch4-window8-256 on an unknown dataset. It achieves the following results on the evaluation set: * Loss: 0.9637 * Accuracy: 0.6667 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * learning\_rate: 4e-05 * train\_batch\_size: 32 * eval\_batch\_size: 32 * seed: 42 * gradient\_accumulation\_steps: 4 * total\_train\_batch\_size: 128 * optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 * lr\_scheduler\_type: linear * lr\_scheduler\_warmup\_ratio: 0.1 * num\_epochs: 42 ### Training results ### Framework versions * Transformers 4.36.2 * Pytorch 2.1.2+cu118 * Datasets 2.16.1 * Tokenizers 0.15.0	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 4e-05\n* train\\_batch\\_size: 32\n* eval\\_batch\\_size: 32\n* seed: 42\n* gradient\\_accumulation\\_steps: 4\n* total\\_train\\_batch\\_size: 128\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* lr\\_scheduler\\_warmup\\_ratio: 0.1\n* num\\_epochs: 42", "### Training results", "### Framework versions\n\n\n* Transformers 4.36.2\n* Pytorch 2.1.2+cu118\n* Datasets 2.16.1\n* Tokenizers 0.15.0" ]	[ "TAGS\n#transformers #tensorboard #safetensors #swinv2 #image-classification #generated_from_trainer #base_model-microsoft/swinv2-tiny-patch4-window8-256 #license-apache-2.0 #autotrain_compatible #endpoints_compatible #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 4e-05\n* train\\_batch\\_size: 32\n* eval\\_batch\\_size: 32\n* seed: 42\n* gradient\\_accumulation\\_steps: 4\n* total\\_train\\_batch\\_size: 128\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: linear\n* lr\\_scheduler\\_warmup\\_ratio: 0.1\n* num\\_epochs: 42", "### Training results", "### Framework versions\n\n\n* Transformers 4.36.2\n* Pytorch 2.1.2+cu118\n* Datasets 2.16.1\n* Tokenizers 0.15.0" ]
null	transformers	# Clevyby/NSM-delta-7B-Q5_K_S-GGUF This model was converted to GGUF format from [`Nitral-AI/NSM-delta-7B`](https://huggingface.co/Nitral-AI/NSM-delta-7B) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space. Refer to the [original model card](https://huggingface.co/Nitral-AI/NSM-delta-7B) for more details on the model.	{"library_name": "transformers", "tags": ["mergekit", "merge", "llama-cpp", "gguf-my-repo"], "base_model": ["ResplendentAI/Aura_v2_7B", "Nitral-AI/NSM-alpha-7B"]}	Clevyby/NSM-delta-7B-Q5_K_S-GGUF	null	[ "transformers", "gguf", "mergekit", "merge", "llama-cpp", "gguf-my-repo", "base_model:ResplendentAI/Aura_v2_7B", "base_model:Nitral-AI/NSM-alpha-7B", "endpoints_compatible", "region:us" ]	null	2024-04-19T04:25:14+00:00	[]	[]	TAGS #transformers #gguf #mergekit #merge #llama-cpp #gguf-my-repo #base_model-ResplendentAI/Aura_v2_7B #base_model-Nitral-AI/NSM-alpha-7B #endpoints_compatible #region-us	# Clevyby/NSM-delta-7B-Q5_K_S-GGUF This model was converted to GGUF format from 'Nitral-AI/NSM-delta-7B' using URL via the URL's GGUF-my-repo space. Refer to the original model card for more details on the model.	[ "# Clevyby/NSM-delta-7B-Q5_K_S-GGUF\nThis model was converted to GGUF format from 'Nitral-AI/NSM-delta-7B' using URL via the URL's GGUF-my-repo space.\nRefer to the original model card for more details on the model." ]	[ "TAGS\n#transformers #gguf #mergekit #merge #llama-cpp #gguf-my-repo #base_model-ResplendentAI/Aura_v2_7B #base_model-Nitral-AI/NSM-alpha-7B #endpoints_compatible #region-us \n", "# Clevyby/NSM-delta-7B-Q5_K_S-GGUF\nThis model was converted to GGUF format from 'Nitral-AI/NSM-delta-7B' using URL via the URL's GGUF-my-repo space.\nRefer to the original model card for more details on the model." ]
text-classification	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # mistral-reward This model is a fine-tuned version of [damienbenveniste/mistral-supervised](https://huggingface.co/damienbenveniste/mistral-supervised) on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results ### Framework versions - Transformers 4.35.2 - Pytorch 2.1.2 - Datasets 2.15.0 - Tokenizers 0.15.0	{"tags": ["generated_from_trainer"], "base_model": "damienbenveniste/mistral-supervised", "model-index": [{"name": "mistral-reward", "results": []}]}	damienbenveniste/mistral-reward	null	[ "transformers", "safetensors", "mistral", "text-classification", "generated_from_trainer", "base_model:damienbenveniste/mistral-supervised", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T04:28:42+00:00	[]	[]	TAGS #transformers #safetensors #mistral #text-classification #generated_from_trainer #base_model-damienbenveniste/mistral-supervised #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	# mistral-reward This model is a fine-tuned version of damienbenveniste/mistral-supervised on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results ### Framework versions - Transformers 4.35.2 - Pytorch 2.1.2 - Datasets 2.15.0 - Tokenizers 0.15.0	[ "# mistral-reward\n\nThis model is a fine-tuned version of damienbenveniste/mistral-supervised on an unknown dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 5e-05\n- train_batch_size: 8\n- eval_batch_size: 8\n- seed: 42\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- num_epochs: 1", "### Training results", "### Framework versions\n\n- Transformers 4.35.2\n- Pytorch 2.1.2\n- Datasets 2.15.0\n- Tokenizers 0.15.0" ]	[ "TAGS\n#transformers #safetensors #mistral #text-classification #generated_from_trainer #base_model-damienbenveniste/mistral-supervised #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "# mistral-reward\n\nThis model is a fine-tuned version of damienbenveniste/mistral-supervised on an unknown dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 5e-05\n- train_batch_size: 8\n- eval_batch_size: 8\n- seed: 42\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- num_epochs: 1", "### Training results", "### Framework versions\n\n- Transformers 4.35.2\n- Pytorch 2.1.2\n- Datasets 2.15.0\n- Tokenizers 0.15.0" ]
text-generation	transformers	# Clevyby/Nyan-Stunna-7B-Q5_K_S-GGUF This model was converted to GGUF format from [`Nitral-AI/Nyan-Stunna-7B`](https://huggingface.co/Nitral-AI/Nyan-Stunna-7B) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space. Refer to the [original model card](https://huggingface.co/Nitral-AI/Nyan-Stunna-7B) for more details on the model. ### Note: The additional files in this repo is for personal usage using Text Gen Webui with llamacpp_hf.	{"license": "other", "library_name": "transformers", "tags": ["mergekit", "merge", "llama-cpp", "gguf-my-repo"], "base_model": ["arlineka/KittyNyanster-v1", "Nitral-AI/KukulStanta-7B"]}	Clevyby/Nyan-Stunna-7B-Q5_K_S-GGUF	null	[ "transformers", "gguf", "mistral", "text-generation", "mergekit", "merge", "llama-cpp", "gguf-my-repo", "base_model:arlineka/KittyNyanster-v1", "base_model:Nitral-AI/KukulStanta-7B", "license:other", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T04:30:43+00:00	[]	[]	TAGS #transformers #gguf #mistral #text-generation #mergekit #merge #llama-cpp #gguf-my-repo #base_model-arlineka/KittyNyanster-v1 #base_model-Nitral-AI/KukulStanta-7B #license-other #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	# Clevyby/Nyan-Stunna-7B-Q5_K_S-GGUF This model was converted to GGUF format from 'Nitral-AI/Nyan-Stunna-7B' using URL via the URL's GGUF-my-repo space. Refer to the original model card for more details on the model. ### Note: The additional files in this repo is for personal usage using Text Gen Webui with llamacpp_hf.	[ "# Clevyby/Nyan-Stunna-7B-Q5_K_S-GGUF\nThis model was converted to GGUF format from 'Nitral-AI/Nyan-Stunna-7B' using URL via the URL's GGUF-my-repo space.\nRefer to the original model card for more details on the model.", "### Note: \nThe additional files in this repo is for personal usage using Text Gen Webui with llamacpp_hf." ]	[ "TAGS\n#transformers #gguf #mistral #text-generation #mergekit #merge #llama-cpp #gguf-my-repo #base_model-arlineka/KittyNyanster-v1 #base_model-Nitral-AI/KukulStanta-7B #license-other #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "# Clevyby/Nyan-Stunna-7B-Q5_K_S-GGUF\nThis model was converted to GGUF format from 'Nitral-AI/Nyan-Stunna-7B' using URL via the URL's GGUF-my-repo space.\nRefer to the original model card for more details on the model.", "### Note: \nThe additional files in this repo is for personal usage using Text Gen Webui with llamacpp_hf." ]
text-generation	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	OwOOwO/dumbo-krillin62	null	[ "transformers", "safetensors", "llama", "text-generation", "conversational", "arxiv:1910.09700", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T04:32:14+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #llama #text-generation #conversational #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #llama #text-generation #conversational #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
null	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": ["unsloth"]}	Zangs3011/tester_123123	null	[ "transformers", "safetensors", "unsloth", "arxiv:1910.09700", "endpoints_compatible", "region:us" ]	null	2024-04-19T04:34:57+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #unsloth #arxiv-1910.09700 #endpoints_compatible #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #unsloth #arxiv-1910.09700 #endpoints_compatible #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
text-generation	transformers	AI Model Name: Llama 3 70B "Built with Meta Llama 3" https://llama.meta.com/llama3/license/ This is the result of running AutoAWQ to quantize the LLaMA-3 70B model to ~4 bits/parameter. To launch an OpenAI-compatible API endpoint on your Linux server with 2x 3090 or 4090 GPUs: ``` git lfs install git clone https://huggingface.co/catid/cat-llama-3-70b-awq-q128-w4-gemm conda create -n vllm70 python=3.10 -y && conda activate vllm70 pip install -U git+https://github.com/vllm-project/vllm.git python -m vllm.entrypoints.openai.api_server --model cat-llama-3-70b-awq-q128-w4-gemm --tensor-parallel-size 2 --gpu-memory-utilization 0.935 ``` Sadly this barely doesn't fit by ~300MB or so.	{}	catid/cat-llama-3-70b-awq-q128-w4-gemm	null	[ "transformers", "safetensors", "llama", "text-generation", "conversational", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "4-bit", "region:us" ]	null	2024-04-19T04:35:00+00:00	[]	[]	TAGS #transformers #safetensors #llama #text-generation #conversational #autotrain_compatible #endpoints_compatible #text-generation-inference #4-bit #region-us	AI Model Name: Llama 3 70B "Built with Meta Llama 3" URL This is the result of running AutoAWQ to quantize the LLaMA-3 70B model to ~4 bits/parameter. To launch an OpenAI-compatible API endpoint on your Linux server with 2x 3090 or 4090 GPUs: Sadly this barely doesn't fit by ~300MB or so.	[]	[ "TAGS\n#transformers #safetensors #llama #text-generation #conversational #autotrain_compatible #endpoints_compatible #text-generation-inference #4-bit #region-us \n" ]
text-generation	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # gpt-neo-1.3B_LAMA_TREx_finetuning_MAGNET This model is a fine-tuned version of [EleutherAI/gpt-neo-1.3B](https://huggingface.co/EleutherAI/gpt-neo-1.3B) on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 32 - eval_batch_size: 32 - seed: 0 - distributed_type: multi-GPU - num_devices: 4 - total_train_batch_size: 128 - total_eval_batch_size: 128 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 3.0 ### Training results ### Framework versions - Transformers 4.33.2 - Pytorch 1.13.1 - Datasets 2.14.5 - Tokenizers 0.13.3	{"license": "mit", "tags": ["generated_from_trainer"], "base_model": "EleutherAI/gpt-neo-1.3B", "model-index": [{"name": "gpt-neo-1.3B_LAMA_TREx_finetuning_MAGNET", "results": []}]}	KimByeongSu/gpt-neo-1.3B_LAMA_TREx_finetuning_MAGNET	null	[ "transformers", "pytorch", "gpt_neo", "text-generation", "generated_from_trainer", "base_model:EleutherAI/gpt-neo-1.3B", "license:mit", "autotrain_compatible", "endpoints_compatible", "region:us" ]	null	2024-04-19T04:35:54+00:00	[]	[]	TAGS #transformers #pytorch #gpt_neo #text-generation #generated_from_trainer #base_model-EleutherAI/gpt-neo-1.3B #license-mit #autotrain_compatible #endpoints_compatible #region-us	# gpt-neo-1.3B_LAMA_TREx_finetuning_MAGNET This model is a fine-tuned version of EleutherAI/gpt-neo-1.3B on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 32 - eval_batch_size: 32 - seed: 0 - distributed_type: multi-GPU - num_devices: 4 - total_train_batch_size: 128 - total_eval_batch_size: 128 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 3.0 ### Training results ### Framework versions - Transformers 4.33.2 - Pytorch 1.13.1 - Datasets 2.14.5 - Tokenizers 0.13.3	[ "# gpt-neo-1.3B_LAMA_TREx_finetuning_MAGNET\n\nThis model is a fine-tuned version of EleutherAI/gpt-neo-1.3B on an unknown dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 2e-05\n- train_batch_size: 32\n- eval_batch_size: 32\n- seed: 0\n- distributed_type: multi-GPU\n- num_devices: 4\n- total_train_batch_size: 128\n- total_eval_batch_size: 128\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- num_epochs: 3.0", "### Training results", "### Framework versions\n\n- Transformers 4.33.2\n- Pytorch 1.13.1\n- Datasets 2.14.5\n- Tokenizers 0.13.3" ]	[ "TAGS\n#transformers #pytorch #gpt_neo #text-generation #generated_from_trainer #base_model-EleutherAI/gpt-neo-1.3B #license-mit #autotrain_compatible #endpoints_compatible #region-us \n", "# gpt-neo-1.3B_LAMA_TREx_finetuning_MAGNET\n\nThis model is a fine-tuned version of EleutherAI/gpt-neo-1.3B on an unknown dataset.", "## Model description\n\nMore information needed", "## Intended uses & limitations\n\nMore information needed", "## Training and evaluation data\n\nMore information needed", "## Training procedure", "### Training hyperparameters\n\nThe following hyperparameters were used during training:\n- learning_rate: 2e-05\n- train_batch_size: 32\n- eval_batch_size: 32\n- seed: 0\n- distributed_type: multi-GPU\n- num_devices: 4\n- total_train_batch_size: 128\n- total_eval_batch_size: 128\n- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n- lr_scheduler_type: linear\n- num_epochs: 3.0", "### Training results", "### Framework versions\n\n- Transformers 4.33.2\n- Pytorch 1.13.1\n- Datasets 2.14.5\n- Tokenizers 0.13.3" ]
text-generation	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	GalaganKV/Mistral-7B-Instruct-v0.2-MultiTask-v6	null	[ "transformers", "safetensors", "mistral", "text-generation", "conversational", "arxiv:1910.09700", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T04:37:56+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #mistral #text-generation #conversational #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #mistral #text-generation #conversational #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
text-generation	transformers	<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # zephyr-7b-sft-full This model is a fine-tuned version of [imone/Mistral_7B_with_EOT_token](https://huggingface.co/imone/Mistral_7B_with_EOT_token) on the generator dataset. It achieves the following results on the evaluation set: - Loss: 0.9067 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 4 - eval_batch_size: 8 - seed: 42 - distributed_type: multi-GPU - num_devices: 8 - total_train_batch_size: 32 - total_eval_batch_size: 64 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 5 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| \|:-------------:\|:-----:\|:-----:\|:---------------:\| \| 0.0936 \| 1.0 \| 2315 \| 0.9145 \| \| 0.7793 \| 2.0 \| 4630 \| 5.4135 \| \| 0.3835 \| 3.0 \| 6945 \| 3.1220 \| \| 0.0959 \| 4.0 \| 9260 \| 1.1934 \| \| 0.0725 \| 5.0 \| 11575 \| 0.9067 \| ### Framework versions - Transformers 4.38.2 - Pytorch 2.1.2+cu118 - Datasets 2.16.1 - Tokenizers 0.15.2	{"license": "apache-2.0", "tags": ["trl", "sft", "generated_from_trainer"], "datasets": ["generator"], "base_model": "imone/Mistral_7B_with_EOT_token", "model-index": [{"name": "zephyr-7b-sft-full", "results": []}]}	RikkiXu/zephyr-7b-sft-full	null	[ "transformers", "tensorboard", "safetensors", "mistral", "text-generation", "trl", "sft", "generated_from_trainer", "conversational", "dataset:generator", "base_model:imone/Mistral_7B_with_EOT_token", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T04:38:35+00:00	[]	[]	TAGS #transformers #tensorboard #safetensors #mistral #text-generation #trl #sft #generated_from_trainer #conversational #dataset-generator #base_model-imone/Mistral_7B_with_EOT_token #license-apache-2.0 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	zephyr-7b-sft-full ================== This model is a fine-tuned version of imone/Mistral\_7B\_with\_EOT\_token on the generator dataset. It achieves the following results on the evaluation set: * Loss: 0.9067 Model description ----------------- More information needed Intended uses & limitations --------------------------- More information needed Training and evaluation data ---------------------------- More information needed Training procedure ------------------ ### Training hyperparameters The following hyperparameters were used during training: * learning\_rate: 2e-05 * train\_batch\_size: 4 * eval\_batch\_size: 8 * seed: 42 * distributed\_type: multi-GPU * num\_devices: 8 * total\_train\_batch\_size: 32 * total\_eval\_batch\_size: 64 * optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 * lr\_scheduler\_type: cosine * lr\_scheduler\_warmup\_ratio: 0.1 * num\_epochs: 5 ### Training results ### Framework versions * Transformers 4.38.2 * Pytorch 2.1.2+cu118 * Datasets 2.16.1 * Tokenizers 0.15.2	[ "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 2e-05\n* train\\_batch\\_size: 4\n* eval\\_batch\\_size: 8\n* seed: 42\n* distributed\\_type: multi-GPU\n* num\\_devices: 8\n* total\\_train\\_batch\\_size: 32\n* total\\_eval\\_batch\\_size: 64\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: cosine\n* lr\\_scheduler\\_warmup\\_ratio: 0.1\n* num\\_epochs: 5", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* Pytorch 2.1.2+cu118\n* Datasets 2.16.1\n* Tokenizers 0.15.2" ]	[ "TAGS\n#transformers #tensorboard #safetensors #mistral #text-generation #trl #sft #generated_from_trainer #conversational #dataset-generator #base_model-imone/Mistral_7B_with_EOT_token #license-apache-2.0 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "### Training hyperparameters\n\n\nThe following hyperparameters were used during training:\n\n\n* learning\\_rate: 2e-05\n* train\\_batch\\_size: 4\n* eval\\_batch\\_size: 8\n* seed: 42\n* distributed\\_type: multi-GPU\n* num\\_devices: 8\n* total\\_train\\_batch\\_size: 32\n* total\\_eval\\_batch\\_size: 64\n* optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08\n* lr\\_scheduler\\_type: cosine\n* lr\\_scheduler\\_warmup\\_ratio: 0.1\n* num\\_epochs: 5", "### Training results", "### Framework versions\n\n\n* Transformers 4.38.2\n* Pytorch 2.1.2+cu118\n* Datasets 2.16.1\n* Tokenizers 0.15.2" ]
null	transformers	## About <!-- ### quantize_version: 1 --> <!-- ### output_tensor_quantised: 1 --> <!-- ### convert_type: --> <!-- ### vocab_type: --> static quants of https://huggingface.co/allknowingroger/WestLakeLaser-12B-MoE <!-- provided-files --> weighted/imatrix quants seem not to be available (by me) at this time. If they do not show up a week or so after the static ones, I have probably not planned for them. Feel free to request them by opening a Community Discussion. ## Usage If you are unsure how to use GGUF files, refer to one of [TheBloke's READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for more details, including on how to concatenate multi-part files. ## Provided Quants (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) \| Link \| Type \| Size/GB \| Notes \| \|:-----\|:-----\|--------:\|:------\| \| [GGUF](https://huggingface.co/mradermacher/WestLakeLaser-12B-MoE-GGUF/resolve/main/WestLakeLaser-12B-MoE.Q2_K.gguf) \| Q2_K \| 4.9 \| \| \| [GGUF](https://huggingface.co/mradermacher/WestLakeLaser-12B-MoE-GGUF/resolve/main/WestLakeLaser-12B-MoE.IQ3_XS.gguf) \| IQ3_XS \| 5.4 \| \| \| [GGUF](https://huggingface.co/mradermacher/WestLakeLaser-12B-MoE-GGUF/resolve/main/WestLakeLaser-12B-MoE.Q3_K_S.gguf) \| Q3_K_S \| 5.7 \| \| \| [GGUF](https://huggingface.co/mradermacher/WestLakeLaser-12B-MoE-GGUF/resolve/main/WestLakeLaser-12B-MoE.IQ3_S.gguf) \| IQ3_S \| 5.7 \| beats Q3_K* \| \| [GGUF](https://huggingface.co/mradermacher/WestLakeLaser-12B-MoE-GGUF/resolve/main/WestLakeLaser-12B-MoE.IQ3_M.gguf) \| IQ3_M \| 5.8 \| \| \| [GGUF](https://huggingface.co/mradermacher/WestLakeLaser-12B-MoE-GGUF/resolve/main/WestLakeLaser-12B-MoE.Q3_K_M.gguf) \| Q3_K_M \| 6.3 \| lower quality \| \| [GGUF](https://huggingface.co/mradermacher/WestLakeLaser-12B-MoE-GGUF/resolve/main/WestLakeLaser-12B-MoE.Q3_K_L.gguf) \| Q3_K_L \| 6.8 \| \| \| [GGUF](https://huggingface.co/mradermacher/WestLakeLaser-12B-MoE-GGUF/resolve/main/WestLakeLaser-12B-MoE.IQ4_XS.gguf) \| IQ4_XS \| 7.1 \| \| \| [GGUF](https://huggingface.co/mradermacher/WestLakeLaser-12B-MoE-GGUF/resolve/main/WestLakeLaser-12B-MoE.Q4_K_S.gguf) \| Q4_K_S \| 7.4 \| fast, recommended \| \| [GGUF](https://huggingface.co/mradermacher/WestLakeLaser-12B-MoE-GGUF/resolve/main/WestLakeLaser-12B-MoE.Q4_K_M.gguf) \| Q4_K_M \| 7.9 \| fast, recommended \| \| [GGUF](https://huggingface.co/mradermacher/WestLakeLaser-12B-MoE-GGUF/resolve/main/WestLakeLaser-12B-MoE.Q5_K_S.gguf) \| Q5_K_S \| 9.0 \| \| \| [GGUF](https://huggingface.co/mradermacher/WestLakeLaser-12B-MoE-GGUF/resolve/main/WestLakeLaser-12B-MoE.Q5_K_M.gguf) \| Q5_K_M \| 9.2 \| \| \| [GGUF](https://huggingface.co/mradermacher/WestLakeLaser-12B-MoE-GGUF/resolve/main/WestLakeLaser-12B-MoE.Q6_K.gguf) \| Q6_K \| 10.7 \| very good quality \| \| [GGUF](https://huggingface.co/mradermacher/WestLakeLaser-12B-MoE-GGUF/resolve/main/WestLakeLaser-12B-MoE.Q8_0.gguf) \| Q8_0 \| 13.8 \| fast, best quality \| Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better): ![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png) And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9 ## FAQ / Model Request See https://huggingface.co/mradermacher/model_requests for some answers to questions you might have and/or if you want some other model quantized. ## Thanks I thank my company, [nethype GmbH](https://www.nethype.de/), for letting me use its servers and providing upgrades to my workstation to enable this work in my free time. <!-- end -->	{"language": ["en"], "license": "apache-2.0", "library_name": "transformers", "tags": ["moe", "frankenmoe", "merge", "mergekit", "lazymergekit", "allknowingroger/PrometheusLaser-7B-slerp", "senseable/WestLake-7B-v2"], "base_model": "allknowingroger/WestLakeLaser-12B-MoE", "quantized_by": "mradermacher"}	mradermacher/WestLakeLaser-12B-MoE-GGUF	null	[ "transformers", "gguf", "moe", "frankenmoe", "merge", "mergekit", "lazymergekit", "allknowingroger/PrometheusLaser-7B-slerp", "senseable/WestLake-7B-v2", "en", "base_model:allknowingroger/WestLakeLaser-12B-MoE", "license:apache-2.0", "endpoints_compatible", "region:us" ]	null	2024-04-19T04:38:59+00:00	[]	[ "en" ]	TAGS #transformers #gguf #moe #frankenmoe #merge #mergekit #lazymergekit #allknowingroger/PrometheusLaser-7B-slerp #senseable/WestLake-7B-v2 #en #base_model-allknowingroger/WestLakeLaser-12B-MoE #license-apache-2.0 #endpoints_compatible #region-us	About ----- static quants of URL weighted/imatrix quants seem not to be available (by me) at this time. If they do not show up a week or so after the static ones, I have probably not planned for them. Feel free to request them by opening a Community Discussion. Usage ----- If you are unsure how to use GGUF files, refer to one of TheBloke's READMEs for more details, including on how to concatenate multi-part files. Provided Quants --------------- (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better): !URL And here are Artefact2's thoughts on the matter: URL FAQ / Model Request ------------------- See URL for some answers to questions you might have and/or if you want some other model quantized. Thanks ------ I thank my company, nethype GmbH, for letting me use its servers and providing upgrades to my workstation to enable this work in my free time.	[]	[ "TAGS\n#transformers #gguf #moe #frankenmoe #merge #mergekit #lazymergekit #allknowingroger/PrometheusLaser-7B-slerp #senseable/WestLake-7B-v2 #en #base_model-allknowingroger/WestLakeLaser-12B-MoE #license-apache-2.0 #endpoints_compatible #region-us \n" ]
null	transformers	# Uploaded model - Developed by: lilyyellow - License: apache-2.0 - Finetuned from model : unsloth/mistral-7b-bnb-4bit This mistral model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library. [<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)	{"language": ["en"], "license": "apache-2.0", "tags": ["text-generation-inference", "transformers", "unsloth", "mistral", "trl"], "base_model": "unsloth/mistral-7b-bnb-4bit"}	lilyyellow/unsloth-mistral-7b-bnb-4bit-sft-add-start-end-index-epoch-2	null	[ "transformers", "safetensors", "text-generation-inference", "unsloth", "mistral", "trl", "en", "base_model:unsloth/mistral-7b-bnb-4bit", "license:apache-2.0", "endpoints_compatible", "region:us" ]	null	2024-04-19T04:41:35+00:00	[]	[ "en" ]	TAGS #transformers #safetensors #text-generation-inference #unsloth #mistral #trl #en #base_model-unsloth/mistral-7b-bnb-4bit #license-apache-2.0 #endpoints_compatible #region-us	# Uploaded model - Developed by: lilyyellow - License: apache-2.0 - Finetuned from model : unsloth/mistral-7b-bnb-4bit This mistral model was trained 2x faster with Unsloth and Huggingface's TRL library. <img src="URL width="200"/>	[ "# Uploaded model\n\n- Developed by: lilyyellow\n- License: apache-2.0\n- Finetuned from model : unsloth/mistral-7b-bnb-4bit\n\nThis mistral model was trained 2x faster with Unsloth and Huggingface's TRL library.\n\n<img src=\"URL width=\"200\"/>" ]	[ "TAGS\n#transformers #safetensors #text-generation-inference #unsloth #mistral #trl #en #base_model-unsloth/mistral-7b-bnb-4bit #license-apache-2.0 #endpoints_compatible #region-us \n", "# Uploaded model\n\n- Developed by: lilyyellow\n- License: apache-2.0\n- Finetuned from model : unsloth/mistral-7b-bnb-4bit\n\nThis mistral model was trained 2x faster with Unsloth and Huggingface's TRL library.\n\n<img src=\"URL width=\"200\"/>" ]
text-generation	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	Iyan/2024-04-19	null	[ "transformers", "safetensors", "mistral", "text-generation", "arxiv:1910.09700", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T04:42:25+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #mistral #text-generation #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #mistral #text-generation #arxiv-1910.09700 #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
text-to-image	diffusers	<!-- This model card has been generated automatically according to the information the training script had access to. You should probably proofread and complete it, then remove this comment. --> # Critical Dream - cosmicBboy/stable-diffusion-xl-base-1.0-lora-dreambooth-critdream-v0.6.0 <Gallery /> ## Model description These are cosmicBboy/stable-diffusion-xl-base-1.0-lora-dreambooth-critdream-v0.6.0 LoRA adaption weights for stabilityai/stable-diffusion-xl-base-1.0, for the purposes of generating images for the [Critical Dream](https://github.com/cosmicBboy/critical-dream) project. The weights were trained using [DreamBooth](https://dreambooth.github.io/). LoRA for the text encoder was enabled: True. Special VAE used for training: stabilityai/sdxl-vae. ## Trigger words You should use a picture of [dm-matt-mercer], a dungeon master. background is a forest. fantasy art style, high quality, highly detailed, sharp focus" to trigger the image generation. ## Download model Weights for this model are available in Safetensors format. [Download](cosmicBboy/stable-diffusion-xl-base-1.0-lora-dreambooth-critdream-v0.6.0/tree/main) them in the Files & versions tab. ## Tracker run link https://wandb.ai/nielsbantilan/dreambooth-lora-sd-xl/runs/9d8tqb1s ## Intended uses & limitations #### How to use ```python # TODO: add an example code snippet for running this diffusion pipeline ``` #### Limitations and bias [TODO: provide examples of latent issues and potential remediations] ## Training details [TODO: describe the data used to train the model]	{"license": "openrail++", "library_name": "diffusers", "tags": ["text-to-image", "stable-diffusion-xl", "stable-diffusion-xl-diffusers", "text-to-image", "diffusers", "lora", "template:sd-lora"], "base_model": "stabilityai/stable-diffusion-xl-base-1.0", "prompt": "a picture of [dm-matt-mercer], a dungeon master. background is a forest. fantasy art style, high quality, highly detailed, sharp focus\"", "widget": [{"text": "a picture of [dm-matt-mercer]", "output": {"url": "image_0.png"}}, {"text": "a picture of [dm-matt-mercer]", "output": {"url": "image_1.png"}}, {"text": "a picture of a dungeon master.", "output": {"url": "image_2.png"}}, {"text": "a picture of a dungeon master.", "output": {"url": "image_3.png"}}, {"text": "a picture of [critrole-fjord], a male half-orc warlock. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_4.png"}}, {"text": "a picture of [critrole-fjord], a male half-orc warlock. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_5.png"}}, {"text": "a picture of a male half-orc warlock", "output": {"url": "image_6.png"}}, {"text": "a picture of a male half-orc warlock", "output": {"url": "image_7.png"}}, {"text": "a picture of [critrole-beau], a female human monk. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_8.png"}}, {"text": "a picture of [critrole-beau], a female human monk. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_9.png"}}, {"text": "a picture of a female human monk", "output": {"url": "image_10.png"}}, {"text": "a picture of a female human monk", "output": {"url": "image_11.png"}}, {"text": "a picture of [critrole-caduceus], a male firbolg cleric. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_12.png"}}, {"text": "a picture of [critrole-caduceus], a male firbolg cleric. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_13.png"}}, {"text": "a picture of a male firbolg cleric", "output": {"url": "image_14.png"}}, {"text": "a picture of a male firbolg cleric", "output": {"url": "image_15.png"}}, {"text": "a picture of [critrole-caleb], a male human wizard. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_16.png"}}, {"text": "a picture of [critrole-caleb], a male human wizard. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_17.png"}}, {"text": "a picture of a male human wizard", "output": {"url": "image_18.png"}}, {"text": "a picture of a male human wizard", "output": {"url": "image_19.png"}}, {"text": "a picture of [critrole-jester], a female tiefling cleric. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_20.png"}}, {"text": "a picture of [critrole-jester], a female tiefling cleric. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_21.png"}}, {"text": "a picture of a female tiefling cleric", "output": {"url": "image_22.png"}}, {"text": "a picture of a female tiefling cleric", "output": {"url": "image_23.png"}}, {"text": "a picture of [critrole-nott], a female goblin rogue. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_24.png"}}, {"text": "a picture of [critrole-nott], a female goblin rogue. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_25.png"}}, {"text": "a picture of a female goblin rogue", "output": {"url": "image_26.png"}}, {"text": "a picture of a female goblin rogue", "output": {"url": "image_27.png"}}, {"text": "a picture of [critrole-veth], a female halfling rogue/wizard. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_28.png"}}, {"text": "a picture of [critrole-veth], a female halfling rogue/wizard. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_29.png"}}, {"text": "a picture of a female halfling rogue/wizard", "output": {"url": "image_30.png"}}, {"text": "a picture of a female halfling rogue/wizard", "output": {"url": "image_31.png"}}, {"text": "a picture of [critrole-yasha], a female aasimar barbarian. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_32.png"}}, {"text": "a picture of [critrole-yasha], a female aasimar barbarian. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_33.png"}}, {"text": "a picture of a female aasimar barbarian", "output": {"url": "image_34.png"}}, {"text": "a picture of a female aasimar barbarian", "output": {"url": "image_35.png"}}, {"text": "a picture of [critrole-mollymauk], a male tiefling blood hunter. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_36.png"}}, {"text": "a picture of [critrole-mollymauk], a male tiefling blood hunter. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_37.png"}}, {"text": "a picture of a male tiefling blood hunter", "output": {"url": "image_38.png"}}, {"text": "a picture of a male tiefling blood hunter", "output": {"url": "image_39.png"}}, {"text": "a picture of [critrole-essek], a male drow wizard. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_40.png"}}, {"text": "a picture of [critrole-essek], a male drow wizard. background is a forest. fantasy art style, high quality, highly detailed, sharp focus", "output": {"url": "image_41.png"}}, {"text": "a picture of a male drow wizard", "output": {"url": "image_42.png"}}, {"text": "a picture of a male drow wizard", "output": {"url": "image_43.png"}}]}	cosmicBboy/stable-diffusion-xl-base-1.0-lora-dreambooth-critdream-v0.6.0	null	[ "diffusers", "text-to-image", "stable-diffusion-xl", "stable-diffusion-xl-diffusers", "lora", "template:sd-lora", "base_model:stabilityai/stable-diffusion-xl-base-1.0", "license:openrail++", "region:us" ]	null	2024-04-19T04:42:37+00:00	[]	[]	TAGS #diffusers #text-to-image #stable-diffusion-xl #stable-diffusion-xl-diffusers #lora #template-sd-lora #base_model-stabilityai/stable-diffusion-xl-base-1.0 #license-openrail++ #region-us	# Critical Dream - cosmicBboy/stable-diffusion-xl-base-1.0-lora-dreambooth-critdream-v0.6.0 <Gallery /> ## Model description These are cosmicBboy/stable-diffusion-xl-base-1.0-lora-dreambooth-critdream-v0.6.0 LoRA adaption weights for stabilityai/stable-diffusion-xl-base-1.0, for the purposes of generating images for the Critical Dream project. The weights were trained using DreamBooth. LoRA for the text encoder was enabled: True. Special VAE used for training: stabilityai/sdxl-vae. ## Trigger words You should use a picture of [dm-matt-mercer], a dungeon master. background is a forest. fantasy art style, high quality, highly detailed, sharp focus" to trigger the image generation. ## Download model Weights for this model are available in Safetensors format. Download them in the Files & versions tab. ## Tracker run link URL ## Intended uses & limitations #### How to use #### Limitations and bias [TODO: provide examples of latent issues and potential remediations] ## Training details [TODO: describe the data used to train the model]	[ "# Critical Dream - cosmicBboy/stable-diffusion-xl-base-1.0-lora-dreambooth-critdream-v0.6.0\n\n<Gallery />", "## Model description\n\nThese are cosmicBboy/stable-diffusion-xl-base-1.0-lora-dreambooth-critdream-v0.6.0 LoRA adaption weights for stabilityai/stable-diffusion-xl-base-1.0, for the purposes of\ngenerating images for the Critical Dream\nproject.\n\nThe weights were trained using DreamBooth.\n\nLoRA for the text encoder was enabled: True.\n\nSpecial VAE used for training: stabilityai/sdxl-vae.", "## Trigger words\n\nYou should use a picture of [dm-matt-mercer], a dungeon master. background is a forest. fantasy art style, high quality, highly detailed, sharp focus\" to trigger the image generation.", "## Download model\n\nWeights for this model are available in Safetensors format.\n\nDownload them in the Files & versions tab.", "## Tracker run link\n\nURL", "## Intended uses & limitations", "#### How to use", "#### Limitations and bias\n\n[TODO: provide examples of latent issues and potential remediations]", "## Training details\n\n[TODO: describe the data used to train the model]" ]	[ "TAGS\n#diffusers #text-to-image #stable-diffusion-xl #stable-diffusion-xl-diffusers #lora #template-sd-lora #base_model-stabilityai/stable-diffusion-xl-base-1.0 #license-openrail++ #region-us \n", "# Critical Dream - cosmicBboy/stable-diffusion-xl-base-1.0-lora-dreambooth-critdream-v0.6.0\n\n<Gallery />", "## Model description\n\nThese are cosmicBboy/stable-diffusion-xl-base-1.0-lora-dreambooth-critdream-v0.6.0 LoRA adaption weights for stabilityai/stable-diffusion-xl-base-1.0, for the purposes of\ngenerating images for the Critical Dream\nproject.\n\nThe weights were trained using DreamBooth.\n\nLoRA for the text encoder was enabled: True.\n\nSpecial VAE used for training: stabilityai/sdxl-vae.", "## Trigger words\n\nYou should use a picture of [dm-matt-mercer], a dungeon master. background is a forest. fantasy art style, high quality, highly detailed, sharp focus\" to trigger the image generation.", "## Download model\n\nWeights for this model are available in Safetensors format.\n\nDownload them in the Files & versions tab.", "## Tracker run link\n\nURL", "## Intended uses & limitations", "#### How to use", "#### Limitations and bias\n\n[TODO: provide examples of latent issues and potential remediations]", "## Training details\n\n[TODO: describe the data used to train the model]" ]
text-generation	transformers	This is the first version of upscaling llama-3. Version 2 is now out and does not have any of the issues that this version has. Please use version 2 instead. Linked bellow: - https://huggingface.co/Replete-AI/Llama-3-11.5B-Instruct-v2 __________________________________________________________________ Llama-3-13B-Instruct Thank you to Meta for the weights for Meta-Llama-3-8B-Instruct ![image/png](https://cdn-uploads.huggingface.co/production/uploads/642cc1c253e76b4c2286c58e/aJJxKus1wP5N-euvHEUq7.png) This is an upscaling of the Meta-Llama-3-8B-Instruct Ai using techniques created for Mistral-Evolved-11b-v0.1. This Ai model has been upscaled from 8b parameters to 13b parameters without any continuous pretraining or fine-tuning. From testing, the model seems to function perfectly at fp16, but has some issues at 4-bit quantization using bitsandbytes. The model that was used to create this one is linked below: https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct	{"license": "other", "library_name": "transformers", "base_model": [], "license_name": "llama-3", "license_link": "https://llama.meta.com/llama3/license/"}	Replete-AI/Llama-3-13B-Instruct	null	[ "transformers", "safetensors", "llama", "text-generation", "conversational", "license:other", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T04:43:16+00:00	[]	[]	TAGS #transformers #safetensors #llama #text-generation #conversational #license-other #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	This is the first version of upscaling llama-3. Version 2 is now out and does not have any of the issues that this version has. Please use version 2 instead. Linked bellow: - URL __________________________________________________________________ Llama-3-13B-Instruct Thank you to Meta for the weights for Meta-Llama-3-8B-Instruct !image/png This is an upscaling of the Meta-Llama-3-8B-Instruct Ai using techniques created for Mistral-Evolved-11b-v0.1. This Ai model has been upscaled from 8b parameters to 13b parameters without any continuous pretraining or fine-tuning. From testing, the model seems to function perfectly at fp16, but has some issues at 4-bit quantization using bitsandbytes. The model that was used to create this one is linked below: URL	[]	[ "TAGS\n#transformers #safetensors #llama #text-generation #conversational #license-other #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n" ]
feature-extraction	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	stvhuang/rcr-run-5pqr6lwp-90396-master-0_20240402T105012-ep23	null	[ "transformers", "safetensors", "xlm-roberta", "feature-extraction", "arxiv:1910.09700", "endpoints_compatible", "region:us" ]	null	2024-04-19T04:44:49+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #xlm-roberta #feature-extraction #arxiv-1910.09700 #endpoints_compatible #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #xlm-roberta #feature-extraction #arxiv-1910.09700 #endpoints_compatible #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
text-to-audio	transformers	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed]	{"library_name": "transformers", "tags": []}	procit001/female_dutch_english_voice_v2	null	[ "transformers", "safetensors", "vits", "text-to-audio", "arxiv:1910.09700", "endpoints_compatible", "region:us" ]	null	2024-04-19T04:46:06+00:00	[ "1910.09700" ]	[]	TAGS #transformers #safetensors #vits #text-to-audio #arxiv-1910.09700 #endpoints_compatible #region-us	# Model Card for Model ID ## Model Details ### Model Description This is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated. - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]	[ "TAGS\n#transformers #safetensors #vits #text-to-audio #arxiv-1910.09700 #endpoints_compatible #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\nThis is the model card of a transformers model that has been pushed on the Hub. This model card has been automatically generated.\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact" ]
text-generation	transformers	# Llama-3-Orca-1.0-8B <!-- Provide a quick summary of what the model is/does. --> ![image/png](https://cdn-uploads.huggingface.co/production/uploads/6437292ecd93f4c9a34b0d47/6XQuhjWNr6C4RbU9f1k99.png) ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> I fine-tuned llama-3 8B on mainly SlimOrca, along with other datasets to improve performance in math, coding, and writing. - Developed by: Locutusque - Model type: Built with Meta Llama 3 - Language(s) (NLP): Many? - License: Llama 3 license https://huggingface.co/meta-llama/Meta-Llama-3-8B/blob/main/LICENSE ## Quants ### EXL2 [@bartowski](https://huggingface.co/bartowski/) - https://huggingface.co/bartowski/Llama-3-Orca-1.0-8B-exl2 ### GGUF [@bartowski](https://huggingface.co/bartowski/) - https://huggingface.co/bartowski/Llama-3-Orca-1.0-8B-GGUF ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> This model has great performance in writing and coding. ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> Conversational AI.	{"license": "other", "library_name": "transformers", "datasets": ["Open-Orca/SlimOrca-Dedup", "jondurbin/airoboros-3.2", "microsoft/orca-math-word-problems-200k", "m-a-p/Code-Feedback", "MaziyarPanahi/WizardLM_evol_instruct_V2_196k"]}	Locutusque/Llama-3-Orca-1.0-8B	null	[ "transformers", "safetensors", "llama", "text-generation", "conversational", "dataset:Open-Orca/SlimOrca-Dedup", "dataset:jondurbin/airoboros-3.2", "dataset:microsoft/orca-math-word-problems-200k", "dataset:m-a-p/Code-Feedback", "dataset:MaziyarPanahi/WizardLM_evol_instruct_V2_196k", "license:other", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T04:49:06+00:00	[]	[]	TAGS #transformers #safetensors #llama #text-generation #conversational #dataset-Open-Orca/SlimOrca-Dedup #dataset-jondurbin/airoboros-3.2 #dataset-microsoft/orca-math-word-problems-200k #dataset-m-a-p/Code-Feedback #dataset-MaziyarPanahi/WizardLM_evol_instruct_V2_196k #license-other #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	# Llama-3-Orca-1.0-8B !image/png ## Model Details ### Model Description I fine-tuned llama-3 8B on mainly SlimOrca, along with other datasets to improve performance in math, coding, and writing. - Developed by: Locutusque - Model type: Built with Meta Llama 3 - Language(s) (NLP): Many? - License: Llama 3 license URL ## Quants ### EXL2 @bartowski - URL ### GGUF @bartowski - URL ## Uses This model has great performance in writing and coding. ### Direct Use Conversational AI.	[ "# Llama-3-Orca-1.0-8B\n\n\n\n\n!image/png", "## Model Details", "### Model Description\n\n\n\nI fine-tuned llama-3 8B on mainly SlimOrca, along with other datasets to improve performance in math, coding, and writing.\n\n- Developed by: Locutusque\n- Model type: Built with Meta Llama 3\n- Language(s) (NLP): Many?\n- License: Llama 3 license URL", "## Quants", "### EXL2 @bartowski\n\n- URL", "### GGUF @bartowski\n\n- URL", "## Uses\n\n\n\nThis model has great performance in writing and coding.", "### Direct Use\n\n\n\nConversational AI." ]	[ "TAGS\n#transformers #safetensors #llama #text-generation #conversational #dataset-Open-Orca/SlimOrca-Dedup #dataset-jondurbin/airoboros-3.2 #dataset-microsoft/orca-math-word-problems-200k #dataset-m-a-p/Code-Feedback #dataset-MaziyarPanahi/WizardLM_evol_instruct_V2_196k #license-other #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "# Llama-3-Orca-1.0-8B\n\n\n\n\n!image/png", "## Model Details", "### Model Description\n\n\n\nI fine-tuned llama-3 8B on mainly SlimOrca, along with other datasets to improve performance in math, coding, and writing.\n\n- Developed by: Locutusque\n- Model type: Built with Meta Llama 3\n- Language(s) (NLP): Many?\n- License: Llama 3 license URL", "## Quants", "### EXL2 @bartowski\n\n- URL", "### GGUF @bartowski\n\n- URL", "## Uses\n\n\n\nThis model has great performance in writing and coding.", "### Direct Use\n\n\n\nConversational AI." ]
null	null	KVQuant is a methodology for efficient KV cache quantization that incorporates several innovations to acheive accurate low-precision quantization, thereby enabling efficient long context length inference. TLDR: KVQuant addresses the memory bottleneck with long context length inference by quantizing the KV cache to low precision. KVQuant achieves high accuracy with low-precision KV cache quantization by considering several consistent patterns observed in cached KV values across different LLMs, and by developing methods to exploit these patterns, including: - Per-channel, Pre-RoPE Key quantization to better match the outlier channels in Keys - Non-Uniform Quantization (NUQ) to better represent the non-uniform activations - Dense-and-Sparse Quantization to mitigate the impacts of numerical outliers on quantization difficulty - Q-Norm to mitigate distribution shift at ultra low precisions (eg. 2-bit) - Attention-Sink Aware Quantization to avoid quantization error with the first token, which is disproportionately sensitive to quantization error For more details please check out our [paper](https://arxiv.org/abs/2401.18079.pdf). ## Model description Quantizer file for running DBRX with 4-bit KV cache using KVQuant. * Base Model: [DBRX](https://www.databricks.com/blog/introducing-dbrx-new-state-art-open-llm) * Bitwidth: 4-bit * Sparsity Level: 1% ## Links * Paper: [https://arxiv.org/abs/2401.18079.pdf](https://arxiv.org/abs/2401.18079.pdf) * Code: [https://github.com/SqueezeAILab/KVQuant](https://github.com/SqueezeAILab/KVQuant) --- license: mit ---	{}	squeeze-ai-lab/dbrx-base-a4-s1	null	[ "arxiv:2401.18079", "region:us" ]	null	2024-04-19T04:49:46+00:00	[ "2401.18079" ]	[]	TAGS #arxiv-2401.18079 #region-us	KVQuant is a methodology for efficient KV cache quantization that incorporates several innovations to acheive accurate low-precision quantization, thereby enabling efficient long context length inference. TLDR: KVQuant addresses the memory bottleneck with long context length inference by quantizing the KV cache to low precision. KVQuant achieves high accuracy with low-precision KV cache quantization by considering several consistent patterns observed in cached KV values across different LLMs, and by developing methods to exploit these patterns, including: - Per-channel, Pre-RoPE Key quantization to better match the outlier channels in Keys - Non-Uniform Quantization (NUQ) to better represent the non-uniform activations - Dense-and-Sparse Quantization to mitigate the impacts of numerical outliers on quantization difficulty - Q-Norm to mitigate distribution shift at ultra low precisions (eg. 2-bit) - Attention-Sink Aware Quantization to avoid quantization error with the first token, which is disproportionately sensitive to quantization error For more details please check out our paper. ## Model description Quantizer file for running DBRX with 4-bit KV cache using KVQuant. * Base Model: DBRX * Bitwidth: 4-bit * Sparsity Level: 1% ## Links * Paper: URL * Code: URL --- license: mit ---	[ "## Model description\n\nQuantizer file for running DBRX with 4-bit KV cache using KVQuant. \n\n* Base Model: DBRX\n* Bitwidth: 4-bit\n* Sparsity Level: 1%", "## Links\n\n* Paper: URL\n* Code: URL \n\n---\nlicense: mit\n---" ]	[ "TAGS\n#arxiv-2401.18079 #region-us \n", "## Model description\n\nQuantizer file for running DBRX with 4-bit KV cache using KVQuant. \n\n* Base Model: DBRX\n* Bitwidth: 4-bit\n* Sparsity Level: 1%", "## Links\n\n* Paper: URL\n* Code: URL \n\n---\nlicense: mit\n---" ]
reinforcement-learning	null	# Reinforce Agent playing CartPole-v1 This is a trained model of a Reinforce agent playing CartPole-v1 . To learn to use this model and train yours check Unit 4 of the Deep Reinforcement Learning Course: https://huggingface.co/deep-rl-course/unit4/introduction	{"tags": ["CartPole-v1", "reinforce", "reinforcement-learning", "custom-implementation", "deep-rl-class"], "model-index": [{"name": "CartPole-v1", "results": [{"task": {"type": "reinforcement-learning", "name": "reinforcement-learning"}, "dataset": {"name": "CartPole-v1", "type": "CartPole-v1"}, "metrics": [{"type": "mean_reward", "value": "500.00 +/- 0.00", "name": "mean_reward", "verified": false}]}]}]}	kelvinho8/CartPole-v1	null	[ "CartPole-v1", "reinforce", "reinforcement-learning", "custom-implementation", "deep-rl-class", "model-index", "region:us" ]	null	2024-04-19T04:52:21+00:00	[]	[]	TAGS #CartPole-v1 #reinforce #reinforcement-learning #custom-implementation #deep-rl-class #model-index #region-us	# Reinforce Agent playing CartPole-v1 This is a trained model of a Reinforce agent playing CartPole-v1 . To learn to use this model and train yours check Unit 4 of the Deep Reinforcement Learning Course: URL	[ "# Reinforce Agent playing CartPole-v1\n This is a trained model of a Reinforce agent playing CartPole-v1 .\n To learn to use this model and train yours check Unit 4 of the Deep Reinforcement Learning Course: URL" ]	[ "TAGS\n#CartPole-v1 #reinforce #reinforcement-learning #custom-implementation #deep-rl-class #model-index #region-us \n", "# Reinforce Agent playing CartPole-v1\n This is a trained model of a Reinforce agent playing CartPole-v1 .\n To learn to use this model and train yours check Unit 4 of the Deep Reinforcement Learning Course: URL" ]
text-generation	transformers	<!-- header start --> <p align="center"> <img src="https://i.imgur.com/mNM6Cai.png" width="100%" alt="Friendli Logo"> </p> <!-- header end --> # Llama 3 70B - FP8 - Model creator: [Meta Llama 3](https://huggingface.co/meta-llama) - Original model: [Llama 3 70B](https://huggingface.co/meta-llama/Meta-Llama-3-70B) ## Description This repo contains the Llama 3 70B model quantized to FP8 by FriendliAI, significantly enhancing its inference efficiency while maintaining high accuracy. Note that FP8 is only supported by NVIDIA Ada, Hopper, and Blackwell GPU architectures. Check out [FriendliAI documentation](https://docs.friendli.ai/) for more details. ## License Refer to the license of the original model card. ## Compatibility This model is compatible with [Friendli Container](https://friendli.ai/products/container/). ## Prerequisites - Before you begin, make sure you have signed up for [Friendli Suite](https://suite.friendli.ai/). You can use Friendli Containers free of charge for four weeks. - Prepare a Personal Access Token following [this guide](#preparing-personal-access-token). - Prepare a Friendli Container Secret following [this guide](#preparing-container-secret). ### Preparing Personal Access Token PAT (Personal Access Token) is the user credential for for logging into our container registry. 1. Sign in [Friendli Suite](https://suite.friendli.ai/). 2. Go to [User Settings > Tokens](https://suite.friendli.ai/user-settings/tokens) and click 'Create new token'. 3. Save your created token value. ### Preparing Container Secret Container secret is a credential to launch our Friendli Container images. You should pass the container secret as an environment variable to run the container image. 1. Sign in [Friendli Suite](https://suite.friendli.ai/). 2. Go to Container > Container Secrets and click 'Create secret'. 3. Save your created secret value. ### Pulling Friendli Container Image 1. Log in to the Docker client using the personal access token created as outlined in [this guide](#preparing-personal-access-token). ```sh export FRIENDLI_PAT="YOUR PAT" docker login registry.friendli.ai -u $YOUR_EMAIL -p $FRIENDLI_PAT ``` 2. Pull image ```sh docker pull registry.friendli.ai/trial ``` ## Running Friendli Container Once you've prepared the image of Friendli Container, you can launch it to create a serving endpoint. ```sh docker run \ --gpus '"device=0,1"' \ -p 8000:8000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ -e FRIENDLI_CONTAINER_SECRET="YOUR CONTAINER SECRET" \ registry.friendli.ai/trial \ --web-server-port 8000 \ --hf-model-name FriendliAI/Meta-Llama-3-70B-fp8 \ --num-devices 2 # Use tensor parallelism degree 2 ``` --- # Original model card: Meta Llama 3 70B ## Model Details Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety. Model developers Meta Variations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants. Input Models input text only. Output Models generate text and code only. Model Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. <table> <tr> <td> </td> <td><strong>Training Data</strong> </td> <td><strong>Params</strong> </td> <td><strong>Context length</strong> </td> <td><strong>GQA</strong> </td> <td><strong>Token count</strong> </td> <td><strong>Knowledge cutoff</strong> </td> </tr> <tr> <td rowspan="2" >Llama 3 </td> <td rowspan="2" >A new mix of publicly available online data. </td> <td>8B </td> <td>8k </td> <td>Yes </td> <td rowspan="2" >15T+ </td> <td>March, 2023 </td> </tr> <tr> <td>70B </td> <td>8k </td> <td>Yes </td> <td>December, 2023 </td> </tr> </table> Llama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability. Model Release Date April 18, 2024. Status This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback. License A custom commercial license is available at: [https://llama.meta.com/llama3/license](https://llama.meta.com/llama3/license) Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model [README](https://github.com/meta-llama/llama3). For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go [here](https://github.com/meta-llama/llama-recipes). ## Intended Use Intended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. Out-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English. Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. ## How to use This repository contains two versions of Meta-Llama-3-8B-Instruct, for use with transformers and with the original `llama3` codebase. ### Use with transformers See the snippet below for usage with Transformers: ```python >>> import transformers >>> import torch >>> model_id = "meta-llama/Meta-Llama-3-70B" >>> pipeline = transformers.pipeline( "text-generation", model=model_id, model_kwargs={"torch_dtype": torch.bfloat16}, device_map="auto" ) >>> pipeline("Hey how are you doing today?") ``` ### Use with `llama3` Please, follow the instructions in the [repository](https://github.com/meta-llama/llama3). To download Original checkpoints, see the example command below leveraging `huggingface-cli`: ``` huggingface-cli download meta-llama/Meta-Llama-3-70B --include "original/" --local-dir Meta-Llama-3-70B ``` For Hugging Face support, we recommend using transformers or TGI, but a similar command works. ## Hardware and Software Training Factors* We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute. Carbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program. <table> <tr> <td> </td> <td><strong>Time (GPU hours)</strong> </td> <td><strong>Power Consumption (W)</strong> </td> <td><strong>Carbon Emitted(tCO2eq)</strong> </td> </tr> <tr> <td>Llama 3 8B </td> <td>1.3M </td> <td>700 </td> <td>390 </td> </tr> <tr> <td>Llama 3 70B </td> <td>6.4M </td> <td>700 </td> <td>1900 </td> </tr> <tr> <td>Total </td> <td>7.7M </td> <td> </td> <td>2290 </td> </tr> </table> CO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others. ## Training Data Overview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data. Data Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively. ## Benchmarks In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see [here](https://github.com/meta-llama/llama3/blob/main/eval_methodology.md). ### Base pretrained models <table> <tr> <td><strong>Category</strong> </td> <td><strong>Benchmark</strong> </td> <td><strong>Llama 3 8B</strong> </td> <td><strong>Llama2 7B</strong> </td> <td><strong>Llama2 13B</strong> </td> <td><strong>Llama 3 70B</strong> </td> <td><strong>Llama2 70B</strong> </td> </tr> <tr> <td rowspan="6" >General </td> <td>MMLU (5-shot) </td> <td>66.6 </td> <td>45.7 </td> <td>53.8 </td> <td>79.5 </td> <td>69.7 </td> </tr> <tr> <td>AGIEval English (3-5 shot) </td> <td>45.9 </td> <td>28.8 </td> <td>38.7 </td> <td>63.0 </td> <td>54.8 </td> </tr> <tr> <td>CommonSenseQA (7-shot) </td> <td>72.6 </td> <td>57.6 </td> <td>67.6 </td> <td>83.8 </td> <td>78.7 </td> </tr> <tr> <td>Winogrande (5-shot) </td> <td>76.1 </td> <td>73.3 </td> <td>75.4 </td> <td>83.1 </td> <td>81.8 </td> </tr> <tr> <td>BIG-Bench Hard (3-shot, CoT) </td> <td>61.1 </td> <td>38.1 </td> <td>47.0 </td> <td>81.3 </td> <td>65.7 </td> </tr> <tr> <td>ARC-Challenge (25-shot) </td> <td>78.6 </td> <td>53.7 </td> <td>67.6 </td> <td>93.0 </td> <td>85.3 </td> </tr> <tr> <td>Knowledge reasoning </td> <td>TriviaQA-Wiki (5-shot) </td> <td>78.5 </td> <td>72.1 </td> <td>79.6 </td> <td>89.7 </td> <td>87.5 </td> </tr> <tr> <td rowspan="4" >Reading comprehension </td> <td>SQuAD (1-shot) </td> <td>76.4 </td> <td>72.2 </td> <td>72.1 </td> <td>85.6 </td> <td>82.6 </td> </tr> <tr> <td>QuAC (1-shot, F1) </td> <td>44.4 </td> <td>39.6 </td> <td>44.9 </td> <td>51.1 </td> <td>49.4 </td> </tr> <tr> <td>BoolQ (0-shot) </td> <td>75.7 </td> <td>65.5 </td> <td>66.9 </td> <td>79.0 </td> <td>73.1 </td> </tr> <tr> <td>DROP (3-shot, F1) </td> <td>58.4 </td> <td>37.9 </td> <td>49.8 </td> <td>79.7 </td> <td>70.2 </td> </tr> </table> ### Instruction tuned models <table> <tr> <td><strong>Benchmark</strong> </td> <td><strong>Llama 3 8B</strong> </td> <td><strong>Llama 2 7B</strong> </td> <td><strong>Llama 2 13B</strong> </td> <td><strong>Llama 3 70B</strong> </td> <td><strong>Llama 2 70B</strong> </td> </tr> <tr> <td>MMLU (5-shot) </td> <td>68.4 </td> <td>34.1 </td> <td>47.8 </td> <td>82.0 </td> <td>52.9 </td> </tr> <tr> <td>GPQA (0-shot) </td> <td>34.2 </td> <td>21.7 </td> <td>22.3 </td> <td>39.5 </td> <td>21.0 </td> </tr> <tr> <td>HumanEval (0-shot) </td> <td>62.2 </td> <td>7.9 </td> <td>14.0 </td> <td>81.7 </td> <td>25.6 </td> </tr> <tr> <td>GSM-8K (8-shot, CoT) </td> <td>79.6 </td> <td>25.7 </td> <td>77.4 </td> <td>93.0 </td> <td>57.5 </td> </tr> <tr> <td>MATH (4-shot, CoT) </td> <td>30.0 </td> <td>3.8 </td> <td>6.7 </td> <td>50.4 </td> <td>11.6 </td> </tr> </table> ### Responsibility & Safety We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community. Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications. Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience. As part of the Llama 3 release, we updated our [Responsible Use Guide](https://llama.meta.com/responsible-use-guide/) to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including [Meta Llama Guard 2](https://llama.meta.com/purple-llama/) and [Code Shield](https://llama.meta.com/purple-llama/) safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a [reference implementation](https://github.com/meta-llama/llama-recipes/tree/main/recipes/responsible_ai) to get you started. #### Llama 3-Instruct As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case. <span style="text-decoration:underline;">Safety</span> For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable. <span style="text-decoration:underline;">Refusals</span> In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2. We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date. #### Responsible release In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision. Misuse If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy/](https://llama.meta.com/llama3/use-policy/). #### Critical risks <span style="text-decoration:underline;">CBRNE</span> (Chemical, Biological, Radiological, Nuclear, and high yield Explosives) We have conducted a two fold assessment of the safety of the model in this area: * Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks. * Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model). ### <span style="text-decoration:underline;">Cyber Security </span> We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of [equivalent coding capability](https://huggingface.co/spaces/facebook/CyberSecEval). ### <span style="text-decoration:underline;">Child Safety</span> Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences. ### Community Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our [Github repository](https://github.com/meta-llama/PurpleLlama). Finally, we put in place a set of resources including an [output reporting mechanism](https://developers.facebook.com/llama_output_feedback) and [bug bounty program](https://www.facebook.com/whitehat) to continuously improve the Llama technology with the help of the community. ## Ethical Considerations and Limitations The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress. But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating [Purple Llama](https://github.com/facebookresearch/PurpleLlama) solutions into your workflows and specifically [Llama Guard](https://ai.meta.com/research/publications/llama-guard-llm-based-input-output-safeguard-for-human-ai-conversations/) which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety. Please see the Responsible Use Guide available at [http://llama.meta.com/responsible-use-guide](http://llama.meta.com/responsible-use-guide) ## Citation instructions @article{llama3modelcard, title={Llama 3 Model Card}, author={AI@Meta}, year={2024}, url = {https://github.com/meta-llama/llama3/blob/main/MODEL_CARD.md} } ## Contributors Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos	{"language": ["en"], "license": "other", "tags": ["facebook", "meta", "pytorch", "llama", "llama-3"], "model_name": "Llama 3 70B", "license_name": "llama3", "license_link": "LICENSE", "extra_gated_prompt": "### META LLAMA 3 COMMUNITY LICENSE AGREEMENT\nMeta Llama 3 Version Release Date: April 18, 2024\n\"Agreement\" means the terms and conditions for use, reproduction, distribution and modification of the Llama Materials set forth herein.\n\"Documentation\" means the specifications, manuals and documentation accompanying Meta Llama 3 distributed by Meta at https://llama.meta.com/get-started/.\n\"Licensee\" or \"you\" means you, or your employer or any other person or entity (if you are entering into this Agreement on such person or entity\u2019s behalf), of the age required under applicable laws, rules or regulations to provide legal consent and that has legal authority to bind your employer or such other person or entity if you are entering in this Agreement on their behalf.\n\"Meta Llama 3\" means the foundational large language models and software and algorithms, including machine-learning model code, trained model weights, inference-enabling code, training-enabling code, fine-tuning enabling code and other elements of the foregoing distributed by Meta at https://llama.meta.com/llama-downloads.\n\"Llama Materials\" means, collectively, Meta\u2019s proprietary Meta Llama 3 and Documentation (and any portion thereof) made available under this Agreement.\n\"Meta\" or \"we\" means Meta Platforms Ireland Limited (if you are located in or, if you are an entity, your principal place of business is in the EEA or Switzerland) and Meta Platforms, Inc. (if you are located outside of the EEA or Switzerland).\n \n1. License Rights and Redistribution.\na. Grant of Rights. You are granted a non-exclusive, worldwide, non-transferable and royalty-free limited license under Meta\u2019s intellectual property or other rights owned by Meta embodied in the Llama Materials to use, reproduce, distribute, copy, create derivative works of, and make modifications to the Llama Materials.\nb. Redistribution and Use.\ni. If you distribute or make available the Llama Materials (or any derivative works thereof), or a product or service that uses any of them, including another AI model, you shall (A) provide a copy of this Agreement with any such Llama Materials; and (B) prominently display \u201cBuilt with Meta Llama 3\u201d on a related website, user interface, blogpost, about page, or product documentation. If you use the Llama Materials to create, train, fine tune, or otherwise improve an AI model, which is distributed or made available, you shall also include \u201cLlama 3\u201d at the beginning of any such AI model name.\nii. If you receive Llama Materials, or any derivative works thereof, from a Licensee as part of an integrated end user product, then Section 2 of this Agreement will not apply to you.\niii. You must retain in all copies of the Llama Materials that you distribute the following attribution notice within a \u201cNotice\u201d text file distributed as a part of such copies: \u201cMeta Llama 3 is licensed under the Meta Llama 3 Community License, Copyright \u00a9 Meta Platforms, Inc. All Rights Reserved.\u201d\niv. Your use of the Llama Materials must comply with applicable laws and regulations (including trade compliance laws and regulations) and adhere to the Acceptable Use Policy for the Llama Materials (available at https://llama.meta.com/llama3/use-policy), which is hereby incorporated by reference into this Agreement.\nv. You will not use the Llama Materials or any output or results of the Llama Materials to improve any other large language model (excluding Meta Llama 3 or derivative works thereof).\n2. Additional Commercial Terms. If, on the Meta Llama 3 version release date, the monthly active users of the products or services made available by or for Licensee, or Licensee\u2019s affiliates, is greater than 700 million monthly active users in the preceding calendar month, you must request a license from Meta, which Meta may grant to you in its sole discretion, and you are not authorized to exercise any of the rights under this Agreement unless or until Meta otherwise expressly grants you such rights.\n3. Disclaimer of Warranty. UNLESS REQUIRED BY APPLICABLE LAW, THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS THEREFROM ARE PROVIDED ON AN \u201cAS IS\u201d BASIS, WITHOUT WARRANTIES OF ANY KIND, AND META DISCLAIMS ALL WARRANTIES OF ANY KIND, BOTH EXPRESS AND IMPLIED, INCLUDING, WITHOUT LIMITATION, ANY WARRANTIES OF TITLE, NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE. YOU ARE SOLELY RESPONSIBLE FOR DETERMINING THE APPROPRIATENESS OF USING OR REDISTRIBUTING THE LLAMA MATERIALS AND ASSUME ANY RISKS ASSOCIATED WITH YOUR USE OF THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS.\n4. Limitation of Liability. IN NO EVENT WILL META OR ITS AFFILIATES BE LIABLE UNDER ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, TORT, NEGLIGENCE, PRODUCTS LIABILITY, OR OTHERWISE, ARISING OUT OF THIS AGREEMENT, FOR ANY LOST PROFITS OR ANY INDIRECT, SPECIAL, CONSEQUENTIAL, INCIDENTAL, EXEMPLARY OR PUNITIVE DAMAGES, EVEN IF META OR ITS AFFILIATES HAVE BEEN ADVISED OF THE POSSIBILITY OF ANY OF THE FOREGOING.\n5. Intellectual Property.\na. No trademark licenses are granted under this Agreement, and in connection with the Llama Materials, neither Meta nor Licensee may use any name or mark owned by or associated with the other or any of its affiliates, except as required for reasonable and customary use in describing and redistributing the Llama Materials or as set forth in this Section 5(a). Meta hereby grants you a license to use \u201cLlama 3\u201d (the \u201cMark\u201d) solely as required to comply with the last sentence of Section 1.b.i. You will comply with Meta\u2019s brand guidelines (currently accessible at https://about.meta.com/brand/resources/meta/company-brand/ ). All goodwill arising out of your use of the Mark will inure to the benefit of Meta.\nb. Subject to Meta\u2019s ownership of Llama Materials and derivatives made by or for Meta, with respect to any derivative works and modifications of the Llama Materials that are made by you, as between you and Meta, you are and will be the owner of such derivative works and modifications.\nc. If you institute litigation or other proceedings against Meta or any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Llama Materials or Meta Llama 3 outputs or results, or any portion of any of the foregoing, constitutes infringement of intellectual property or other rights owned or licensable by you, then any licenses granted to you under this Agreement shall terminate as of the date such litigation or claim is filed or instituted. You will indemnify and hold harmless Meta from and against any claim by any third party arising out of or related to your use or distribution of the Llama Materials.\n6. Term and Termination. The term of this Agreement will commence upon your acceptance of this Agreement or access to the Llama Materials and will continue in full force and effect until terminated in accordance with the terms and conditions herein. Meta may terminate this Agreement if you are in breach of any term or condition of this Agreement. Upon termination of this Agreement, you shall delete and cease use of the Llama Materials. Sections 3, 4 and 7 shall survive the termination of this Agreement.\n7. Governing Law and Jurisdiction. This Agreement will be governed and construed under the laws of the State of California without regard to choice of law principles, and the UN Convention on Contracts for the International Sale of Goods does not apply to this Agreement. The courts of California shall have exclusive jurisdiction of any dispute arising out of this Agreement.\n### Meta Llama 3 Acceptable Use Policy\nMeta is committed to promoting safe and fair use of its tools and features, including Meta Llama 3. If you access or use Meta Llama 3, you agree to this Acceptable Use Policy (\u201cPolicy\u201d). The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy](https://llama.meta.com/llama3/use-policy)\n#### Prohibited Uses\nWe want everyone to use Meta Llama 3 safely and responsibly. You agree you will not use, or allow others to use, Meta Llama 3 to: 1. Violate the law or others\u2019 rights, including to:\n 1. Engage in, promote, generate, contribute to, encourage, plan, incite, or further illegal or unlawful activity or content, such as:\n 1. Violence or terrorism\n 2. Exploitation or harm to children, including the solicitation, creation, acquisition, or dissemination of child exploitative content or failure to report Child Sexual Abuse Material\n 3. Human trafficking, exploitation, and sexual violence\n 4. The illegal distribution of information or materials to minors, including obscene materials, or failure to employ legally required age-gating in connection with such information or materials.\n 5. Sexual solicitation\n 6. Any other criminal activity\n 2. Engage in, promote, incite, or facilitate the harassment, abuse, threatening, or bullying of individuals or groups of individuals\n 3. Engage in, promote, incite, or facilitate discrimination or other unlawful or harmful conduct in the provision of employment, employment benefits, credit, housing, other economic benefits, or other essential goods and services\n 4. Engage in the unauthorized or unlicensed practice of any profession including, but not limited to, financial, legal, medical/health, or related professional practices\n 5. Collect, process, disclose, generate, or infer health, demographic, or other sensitive personal or private information about individuals without rights and consents required by applicable laws\n 6. Engage in or facilitate any action or generate any content that infringes, misappropriates, or otherwise violates any third-party rights, including the outputs or results of any products or services using the Llama Materials\n 7. Create, generate, or facilitate the creation of malicious code, malware, computer viruses or do anything else that could disable, overburden, interfere with or impair the proper working, integrity, operation or appearance of a website or computer system\n2. Engage in, promote, incite, facilitate, or assist in the planning or development of activities that present a risk of death or bodily harm to individuals, including use of Meta Llama 3 related to the following:\n 1. Military, warfare, nuclear industries or applications, espionage, use for materials or activities that are subject to the International Traffic Arms Regulations (ITAR) maintained by the United States Department of State\n 2. Guns and illegal weapons (including weapon development)\n 3. Illegal drugs and regulated/controlled substances\n 4. Operation of critical infrastructure, transportation technologies, or heavy machinery\n 5. Self-harm or harm to others, including suicide, cutting, and eating disorders\n 6. Any content intended to incite or promote violence, abuse, or any infliction of bodily harm to an individual\n3. Intentionally deceive or mislead others, including use of Meta Llama 3 related to the following:\n 1. Generating, promoting, or furthering fraud or the creation or promotion of disinformation\n 2. Generating, promoting, or furthering defamatory content, including the creation of defamatory statements, images, or other content\n 3. Generating, promoting, or further distributing spam\n 4. Impersonating another individual without consent, authorization, or legal right\n 5. Representing that the use of Meta Llama 3 or outputs are human-generated\n 6. Generating or facilitating false online engagement, including fake reviews and other means of fake online engagement\n4. Fail to appropriately disclose to end users any known dangers of your AI system\nPlease report any violation of this Policy, software \u201cbug,\u201d or other problems that could lead to a violation of this Policy through one of the following means:\n * Reporting issues with the model: [https://github.com/meta-llama/llama3](https://github.com/meta-llama/llama3)\n * Reporting risky content generated by the model:\n developers.facebook.com/llama_output_feedback\n * Reporting bugs and security concerns: facebook.com/whitehat/info\n * Reporting violations of the Acceptable Use Policy or unlicensed uses of Meta Llama 3: [email protected]", "extra_gated_fields": {"First Name": "text", "Last Name": "text", "Date of birth": "date_picker", "Country": "country", "Affiliation": "text", "geo": "ip_location", "By clicking Submit below I accept the terms of the license and acknowledge that the information I provide will be collected stored processed and shared in accordance with the Meta Privacy Policy": "checkbox"}, "extra_gated_description": "The information you provide will be collected, stored, processed and shared in accordance with the [Meta Privacy Policy](https://www.facebook.com/privacy/policy/).", "extra_gated_button_content": "Submit", "base_model": "meta-llama/Meta-Llama-3-70B-Instruct", "inference": false, "model_creator": "Meta Llama 3", "model_link": "https://huggingface.co/meta-llama/Meta-Llama-3-70B", "model_type": "llama", "pipeline_tag": "text-generation", "quantized_by": "FriendliAI"}	FriendliAI/Meta-Llama-3-70B-fp8	null	[ "transformers", "safetensors", "llama", "text-generation", "facebook", "meta", "pytorch", "llama-3", "conversational", "en", "base_model:meta-llama/Meta-Llama-3-70B-Instruct", "license:other", "autotrain_compatible", "text-generation-inference", "8-bit", "region:us" ]	null	2024-04-19T04:52:26+00:00	[]	[ "en" ]	TAGS #transformers #safetensors #llama #text-generation #facebook #meta #pytorch #llama-3 #conversational #en #base_model-meta-llama/Meta-Llama-3-70B-Instruct #license-other #autotrain_compatible #text-generation-inference #8-bit #region-us	![Friendli Logo](https://i.URL width=) Llama 3 70B - FP8 ================= * Model creator: Meta Llama 3 * Original model: Llama 3 70B Description ----------- This repo contains the Llama 3 70B model quantized to FP8 by FriendliAI, significantly enhancing its inference efficiency while maintaining high accuracy. Note that FP8 is only supported by NVIDIA Ada, Hopper, and Blackwell GPU architectures. Check out FriendliAI documentation for more details. License ------- Refer to the license of the original model card. Compatibility ------------- This model is compatible with Friendli Container. Prerequisites ------------- * Before you begin, make sure you have signed up for Friendli Suite. You can use Friendli Containers free of charge for four weeks. * Prepare a Personal Access Token following this guide. * Prepare a Friendli Container Secret following this guide. ### Preparing Personal Access Token PAT (Personal Access Token) is the user credential for for logging into our container registry. 1. Sign in Friendli Suite. 2. Go to User Settings > Tokens and click 'Create new token'. 3. Save your created token value. ### Preparing Container Secret Container secret is a credential to launch our Friendli Container images. You should pass the container secret as an environment variable to run the container image. 1. Sign in Friendli Suite. 2. Go to Container > Container Secrets and click 'Create secret'. 3. Save your created secret value. ### Pulling Friendli Container Image 1. Log in to the Docker client using the personal access token created as outlined in this guide. 2. Pull image Running Friendli Container -------------------------- Once you've prepared the image of Friendli Container, you can launch it to create a serving endpoint. --- Original model card: Meta Llama 3 70B ===================================== Model Details ------------- Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety. Model developers Meta Variations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants. Input Models input text only. Output Models generate text and code only. Model Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. Llama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability. Model Release Date April 18, 2024. Status This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback. License A custom commercial license is available at: URL Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model README. For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go here. Intended Use ------------ Intended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. Out-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English. Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. How to use ---------- This repository contains two versions of Meta-Llama-3-8B-Instruct, for use with transformers and with the original 'llama3' codebase. ### Use with transformers See the snippet below for usage with Transformers: ### Use with 'llama3' Please, follow the instructions in the repository. To download Original checkpoints, see the example command below leveraging 'huggingface-cli': For Hugging Face support, we recommend using transformers or TGI, but a similar command works. Hardware and Software --------------------- Training Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute. Carbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program. CO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others. Training Data ------------- Overview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data. Data Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively. Benchmarks ---------- In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here. ### Base pretrained models ### Instruction tuned models ### Responsibility & Safety We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community. Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications. Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience. As part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started. #### Llama 3-Instruct As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case. Safety For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable. Refusals In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2. We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date. #### Responsible release In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision. Misuse If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL #### Critical risks CBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives) We have conducted a two fold assessment of the safety of the model in this area: * Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks. * Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model). ### Cyber Security We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability. ### Child Safety Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences. ### Community Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository. Finally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community. Ethical Considerations and Limitations -------------------------------------- The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress. But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety. Please see the Responsible Use Guide available at URL instructions @article{llama3modelcard, title={Llama 3 Model Card}, author={AI@Meta}, year={2024}, url = {URL } Contributors ------------ Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos	[ "### Preparing Personal Access Token\n\n\nPAT (Personal Access Token) is the user credential for for logging into our container registry.\n\n\n1. Sign in Friendli Suite.\n2. Go to User Settings > Tokens and click 'Create new token'.\n3. Save your created token value.", "### Preparing Container Secret\n\n\nContainer secret is a credential to launch our Friendli Container images.\nYou should pass the container secret as an environment variable to run the container image.\n\n\n1. Sign in Friendli Suite.\n2. Go to Container > Container Secrets and click 'Create secret'.\n3. Save your created secret value.", "### Pulling Friendli Container Image\n\n\n1. Log in to the Docker client using the personal access token created as outlined in this guide.\n2. Pull image\n\n\nRunning Friendli Container\n--------------------------\n\n\nOnce you've prepared the image of Friendli Container, you can launch it to create a serving endpoint.\n\n\n\n\n---\n\n\nOriginal model card: Meta Llama 3 70B\n=====================================\n\n\nModel Details\n-------------\n\n\nMeta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety.\n\n\nModel developers Meta\n\n\nVariations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants.\n\n\nInput Models input text only.\n\n\nOutput Models generate text and code only.\n\n\nModel Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.\n\n\n\nLlama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability.\n\n\nModel Release Date April 18, 2024.\n\n\nStatus This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback.\n\n\nLicense A custom commercial license is available at: URL\n\n\nWhere to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model README. For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go here.\n\n\nIntended Use\n------------\n\n\nIntended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks.\n\n\nOut-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English.\n\n\nNote: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy.\n\n\nHow to use\n----------\n\n\nThis repository contains two versions of Meta-Llama-3-8B-Instruct, for use with transformers and with the original 'llama3' codebase.", "### Use with transformers\n\n\nSee the snippet below for usage with Transformers:", "### Use with 'llama3'\n\n\nPlease, follow the instructions in the repository.\n\n\nTo download Original checkpoints, see the example command below leveraging 'huggingface-cli':\n\n\nFor Hugging Face support, we recommend using transformers or TGI, but a similar command works.\n\n\nHardware and Software\n---------------------\n\n\nTraining Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.\n\n\nCarbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.\n\n\n\nCO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.\n\n\nTraining Data\n-------------\n\n\nOverview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.\n\n\nData Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.\n\n\nBenchmarks\n----------\n\n\nIn this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here.", "### Base pretrained models", "### Instruction tuned models", "### Responsibility & Safety\n\n\nWe believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.\n\n\nFoundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.\n\n\nRather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.\n\n\nAs part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started.", "#### Llama 3-Instruct\n\n\nAs outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.\n\n\nSafety\n\n\nFor our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.\n\n\nRefusals\n\n\nIn addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.\n\n\nWe built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.", "#### Responsible release\n\n\nIn addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.\n\n\nMisuse\n\n\nIf you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL", "#### Critical risks\n\n\nCBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)\n\n\nWe have conducted a two fold assessment of the safety of the model in this area:\n\n\n* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.\n* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).", "### Cyber Security\n\n\nWe have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability.", "### Child Safety\n\n\nChild Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.", "### Community\n\n\nGenerative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository.\n\n\nFinally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community.\n\n\nEthical Considerations and Limitations\n--------------------------------------\n\n\nThe core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.\n\n\nBut Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.\n\n\nPlease see the Responsible Use Guide available at URL\n\n\ninstructions\n\n\n@article{llama3modelcard,\n\n\ntitle={Llama 3 Model Card},\n\n\nauthor={AI@Meta},\n\n\nyear={2024},\n\n\nurl = {URL\n\n\n}\n\n\nContributors\n------------\n\n\nAaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos" ]	[ "TAGS\n#transformers #safetensors #llama #text-generation #facebook #meta #pytorch #llama-3 #conversational #en #base_model-meta-llama/Meta-Llama-3-70B-Instruct #license-other #autotrain_compatible #text-generation-inference #8-bit #region-us \n", "### Preparing Personal Access Token\n\n\nPAT (Personal Access Token) is the user credential for for logging into our container registry.\n\n\n1. Sign in Friendli Suite.\n2. Go to User Settings > Tokens and click 'Create new token'.\n3. Save your created token value.", "### Preparing Container Secret\n\n\nContainer secret is a credential to launch our Friendli Container images.\nYou should pass the container secret as an environment variable to run the container image.\n\n\n1. Sign in Friendli Suite.\n2. Go to Container > Container Secrets and click 'Create secret'.\n3. Save your created secret value.", "### Pulling Friendli Container Image\n\n\n1. Log in to the Docker client using the personal access token created as outlined in this guide.\n2. Pull image\n\n\nRunning Friendli Container\n--------------------------\n\n\nOnce you've prepared the image of Friendli Container, you can launch it to create a serving endpoint.\n\n\n\n\n---\n\n\nOriginal model card: Meta Llama 3 70B\n=====================================\n\n\nModel Details\n-------------\n\n\nMeta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety.\n\n\nModel developers Meta\n\n\nVariations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants.\n\n\nInput Models input text only.\n\n\nOutput Models generate text and code only.\n\n\nModel Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.\n\n\n\nLlama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability.\n\n\nModel Release Date April 18, 2024.\n\n\nStatus This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback.\n\n\nLicense A custom commercial license is available at: URL\n\n\nWhere to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model README. For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go here.\n\n\nIntended Use\n------------\n\n\nIntended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks.\n\n\nOut-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English.\n\n\nNote: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy.\n\n\nHow to use\n----------\n\n\nThis repository contains two versions of Meta-Llama-3-8B-Instruct, for use with transformers and with the original 'llama3' codebase.", "### Use with transformers\n\n\nSee the snippet below for usage with Transformers:", "### Use with 'llama3'\n\n\nPlease, follow the instructions in the repository.\n\n\nTo download Original checkpoints, see the example command below leveraging 'huggingface-cli':\n\n\nFor Hugging Face support, we recommend using transformers or TGI, but a similar command works.\n\n\nHardware and Software\n---------------------\n\n\nTraining Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.\n\n\nCarbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.\n\n\n\nCO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.\n\n\nTraining Data\n-------------\n\n\nOverview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.\n\n\nData Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.\n\n\nBenchmarks\n----------\n\n\nIn this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here.", "### Base pretrained models", "### Instruction tuned models", "### Responsibility & Safety\n\n\nWe believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.\n\n\nFoundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.\n\n\nRather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.\n\n\nAs part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started.", "#### Llama 3-Instruct\n\n\nAs outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.\n\n\nSafety\n\n\nFor our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.\n\n\nRefusals\n\n\nIn addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.\n\n\nWe built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.", "#### Responsible release\n\n\nIn addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.\n\n\nMisuse\n\n\nIf you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL", "#### Critical risks\n\n\nCBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)\n\n\nWe have conducted a two fold assessment of the safety of the model in this area:\n\n\n* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.\n* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).", "### Cyber Security\n\n\nWe have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability.", "### Child Safety\n\n\nChild Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.", "### Community\n\n\nGenerative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository.\n\n\nFinally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community.\n\n\nEthical Considerations and Limitations\n--------------------------------------\n\n\nThe core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.\n\n\nBut Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.\n\n\nPlease see the Responsible Use Guide available at URL\n\n\ninstructions\n\n\n@article{llama3modelcard,\n\n\ntitle={Llama 3 Model Card},\n\n\nauthor={AI@Meta},\n\n\nyear={2024},\n\n\nurl = {URL\n\n\n}\n\n\nContributors\n------------\n\n\nAaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos" ]
text-generation	transformers	<!-- header start --> <p align="center"> <img src="https://i.imgur.com/mNM6Cai.png" width="100%" alt="Friendli Logo"> </p> <!-- header end --> # Llama 3 70B Instruct - FP8 - Model creator: [Meta Llama 3](https://huggingface.co/meta-llama) - Original model: [Llama 3 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct) ## Description This repo contains the Llama 3 70B Instruct model quantized to FP8 by FriendliAI, significantly enhancing its inference efficiency while maintaining high accuracy. Note that FP8 is only supported by NVIDIA Ada, Hopper, and Blackwell GPU architectures. Check out [FriendliAI documentation](https://docs.friendli.ai/) for more details. ## License Refer to the license of the original model card. ## Compatibility This model is compatible with [Friendli Container](https://friendli.ai/products/container/). ## Prerequisites - Before you begin, make sure you have signed up for [Friendli Suite](https://suite.friendli.ai/). You can use Friendli Containers free of charge for four weeks. - Prepare a Personal Access Token following [this guide](#preparing-personal-access-token). - Prepare a Friendli Container Secret following [this guide](#preparing-container-secret). ### Preparing Personal Access Token PAT (Personal Access Token) is the user credential for for logging into our container registry. 1. Sign in [Friendli Suite](https://suite.friendli.ai/). 2. Go to [User Settings > Tokens](https://suite.friendli.ai/user-settings/tokens) and click 'Create new token'. 3. Save your created token value. ### Preparing Container Secret Container secret is a credential to launch our Friendli Container images. You should pass the container secret as an environment variable to run the container image. 1. Sign in [Friendli Suite](https://suite.friendli.ai/). 2. Go to Container > Container Secrets and click 'Create secret'. 3. Save your created secret value. ### Pulling Friendli Container Image 1. Log in to the Docker client using the personal access token created as outlined in [this guide](#preparing-personal-access-token). ```sh export FRIENDLI_PAT="YOUR PAT" docker login registry.friendli.ai -u $YOUR_EMAIL -p $FRIENDLI_PAT ``` 2. Pull image ```sh docker pull registry.friendli.ai/trial ``` ## Running Friendli Container Once you've prepared the image of Friendli Container, you can launch it to create a serving endpoint. ```sh docker run \ --gpus '"device=0,1"' \ -p 8000:8000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ -e FRIENDLI_CONTAINER_SECRET="YOUR CONTAINER SECRET" \ registry.friendli.ai/trial \ --web-server-port 8000 \ --hf-model-name FriendliAI/Meta-Llama-3-70B-Instruct-fp8 \ --num-devices 2 # Use tensor parallelism degree 2 ``` --- # Original model card: Meta Llama 3 70B Instruct ## Model Details Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety. Model developers Meta Variations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants. Input Models input text only. Output Models generate text and code only. Model Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. <table> <tr> <td> </td> <td><strong>Training Data</strong> </td> <td><strong>Params</strong> </td> <td><strong>Context length</strong> </td> <td><strong>GQA</strong> </td> <td><strong>Token count</strong> </td> <td><strong>Knowledge cutoff</strong> </td> </tr> <tr> <td rowspan="2" >Llama 3 </td> <td rowspan="2" >A new mix of publicly available online data. </td> <td>8B </td> <td>8k </td> <td>Yes </td> <td rowspan="2" >15T+ </td> <td>March, 2023 </td> </tr> <tr> <td>70B </td> <td>8k </td> <td>Yes </td> <td>December, 2023 </td> </tr> </table> Llama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability. Model Release Date April 18, 2024. Status This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback. License A custom commercial license is available at: [https://llama.meta.com/llama3/license](https://llama.meta.com/llama3/license) Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model [README](https://github.com/meta-llama/llama3). For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go [here](https://github.com/meta-llama/llama-recipes). ## Intended Use Intended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. Out-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English. Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. ## How to use This repository contains two versions of Meta-Llama-3-70B-Instruct, for use with transformers and with the original `llama3` codebase. ### Use with transformers See the snippet below for usage with Transformers: ```python import transformers import torch model_id = "meta-llama/Meta-Llama-3-70B-Instruct" pipeline = transformers.pipeline( "text-generation", model=model_id, model_kwargs={"torch_dtype": torch.bfloat16}, device="auto", ) messages = [ {"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"}, {"role": "user", "content": "Who are you?"}, ] prompt = pipeline.tokenizer.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) terminators = [ pipeline.tokenizer.eos_token_id, pipeline.tokenizer.convert_tokens_to_ids("<\|eot_id\|>") ] outputs = pipeline( prompt, max_new_tokens=256, eos_token_id=terminators, do_sample=True, temperature=0.6, top_p=0.9, ) print(outputs[0]["generated_text"][len(prompt):]) ``` ### Use with `llama3` Please, follow the instructions in the [repository](https://github.com/meta-llama/llama3). To download Original checkpoints, see the example command below leveraging `huggingface-cli`: ``` huggingface-cli download meta-llama/Meta-Llama-3-70B-Instruct --include "original/" --local-dir Meta-Llama-3-70B-Instruct ``` For Hugging Face support, we recommend using transformers or TGI, but a similar command works. ## Hardware and Software Training Factors* We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute. Carbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program. <table> <tr> <td> </td> <td><strong>Time (GPU hours)</strong> </td> <td><strong>Power Consumption (W)</strong> </td> <td><strong>Carbon Emitted(tCO2eq)</strong> </td> </tr> <tr> <td>Llama 3 8B </td> <td>1.3M </td> <td>700 </td> <td>390 </td> </tr> <tr> <td>Llama 3 70B </td> <td>6.4M </td> <td>700 </td> <td>1900 </td> </tr> <tr> <td>Total </td> <td>7.7M </td> <td> </td> <td>2290 </td> </tr> </table> CO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others. ## Training Data Overview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data. Data Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively. ## Benchmarks In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see [here](https://github.com/meta-llama/llama3/blob/main/eval_methodology.md). ### Base pretrained models <table> <tr> <td><strong>Category</strong> </td> <td><strong>Benchmark</strong> </td> <td><strong>Llama 3 8B</strong> </td> <td><strong>Llama2 7B</strong> </td> <td><strong>Llama2 13B</strong> </td> <td><strong>Llama 3 70B</strong> </td> <td><strong>Llama2 70B</strong> </td> </tr> <tr> <td rowspan="6" >General </td> <td>MMLU (5-shot) </td> <td>66.6 </td> <td>45.7 </td> <td>53.8 </td> <td>79.5 </td> <td>69.7 </td> </tr> <tr> <td>AGIEval English (3-5 shot) </td> <td>45.9 </td> <td>28.8 </td> <td>38.7 </td> <td>63.0 </td> <td>54.8 </td> </tr> <tr> <td>CommonSenseQA (7-shot) </td> <td>72.6 </td> <td>57.6 </td> <td>67.6 </td> <td>83.8 </td> <td>78.7 </td> </tr> <tr> <td>Winogrande (5-shot) </td> <td>76.1 </td> <td>73.3 </td> <td>75.4 </td> <td>83.1 </td> <td>81.8 </td> </tr> <tr> <td>BIG-Bench Hard (3-shot, CoT) </td> <td>61.1 </td> <td>38.1 </td> <td>47.0 </td> <td>81.3 </td> <td>65.7 </td> </tr> <tr> <td>ARC-Challenge (25-shot) </td> <td>78.6 </td> <td>53.7 </td> <td>67.6 </td> <td>93.0 </td> <td>85.3 </td> </tr> <tr> <td>Knowledge reasoning </td> <td>TriviaQA-Wiki (5-shot) </td> <td>78.5 </td> <td>72.1 </td> <td>79.6 </td> <td>89.7 </td> <td>87.5 </td> </tr> <tr> <td rowspan="4" >Reading comprehension </td> <td>SQuAD (1-shot) </td> <td>76.4 </td> <td>72.2 </td> <td>72.1 </td> <td>85.6 </td> <td>82.6 </td> </tr> <tr> <td>QuAC (1-shot, F1) </td> <td>44.4 </td> <td>39.6 </td> <td>44.9 </td> <td>51.1 </td> <td>49.4 </td> </tr> <tr> <td>BoolQ (0-shot) </td> <td>75.7 </td> <td>65.5 </td> <td>66.9 </td> <td>79.0 </td> <td>73.1 </td> </tr> <tr> <td>DROP (3-shot, F1) </td> <td>58.4 </td> <td>37.9 </td> <td>49.8 </td> <td>79.7 </td> <td>70.2 </td> </tr> </table> ### Instruction tuned models <table> <tr> <td><strong>Benchmark</strong> </td> <td><strong>Llama 3 8B</strong> </td> <td><strong>Llama 2 7B</strong> </td> <td><strong>Llama 2 13B</strong> </td> <td><strong>Llama 3 70B</strong> </td> <td><strong>Llama 2 70B</strong> </td> </tr> <tr> <td>MMLU (5-shot) </td> <td>68.4 </td> <td>34.1 </td> <td>47.8 </td> <td>82.0 </td> <td>52.9 </td> </tr> <tr> <td>GPQA (0-shot) </td> <td>34.2 </td> <td>21.7 </td> <td>22.3 </td> <td>39.5 </td> <td>21.0 </td> </tr> <tr> <td>HumanEval (0-shot) </td> <td>62.2 </td> <td>7.9 </td> <td>14.0 </td> <td>81.7 </td> <td>25.6 </td> </tr> <tr> <td>GSM-8K (8-shot, CoT) </td> <td>79.6 </td> <td>25.7 </td> <td>77.4 </td> <td>93.0 </td> <td>57.5 </td> </tr> <tr> <td>MATH (4-shot, CoT) </td> <td>30.0 </td> <td>3.8 </td> <td>6.7 </td> <td>50.4 </td> <td>11.6 </td> </tr> </table> ### Responsibility & Safety We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community. Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications. Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience. As part of the Llama 3 release, we updated our [Responsible Use Guide](https://llama.meta.com/responsible-use-guide/) to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including [Meta Llama Guard 2](https://llama.meta.com/purple-llama/) and [Code Shield](https://llama.meta.com/purple-llama/) safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a [reference implementation](https://github.com/meta-llama/llama-recipes/tree/main/recipes/responsible_ai) to get you started. #### Llama 3-Instruct As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case. <span style="text-decoration:underline;">Safety</span> For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable. <span style="text-decoration:underline;">Refusals</span> In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2. We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date. #### Responsible release In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision. Misuse If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy/](https://llama.meta.com/llama3/use-policy/). #### Critical risks <span style="text-decoration:underline;">CBRNE</span> (Chemical, Biological, Radiological, Nuclear, and high yield Explosives) We have conducted a two fold assessment of the safety of the model in this area: * Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks. * Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model). ### <span style="text-decoration:underline;">Cyber Security </span> We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of [equivalent coding capability](https://huggingface.co/spaces/facebook/CyberSecEval). ### <span style="text-decoration:underline;">Child Safety</span> Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences. ### Community Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our [Github repository](https://github.com/meta-llama/PurpleLlama). Finally, we put in place a set of resources including an [output reporting mechanism](https://developers.facebook.com/llama_output_feedback) and [bug bounty program](https://www.facebook.com/whitehat) to continuously improve the Llama technology with the help of the community. ## Ethical Considerations and Limitations The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress. But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating [Purple Llama](https://github.com/facebookresearch/PurpleLlama) solutions into your workflows and specifically [Llama Guard](https://ai.meta.com/research/publications/llama-guard-llm-based-input-output-safeguard-for-human-ai-conversations/) which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety. Please see the Responsible Use Guide available at [http://llama.meta.com/responsible-use-guide](http://llama.meta.com/responsible-use-guide) ## Citation instructions @article{llama3modelcard, title={Llama 3 Model Card}, author={AI@Meta}, year={2024}, url = {https://github.com/meta-llama/llama3/blob/main/MODEL_CARD.md} } ## Contributors Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos	{"language": ["en"], "license": "other", "tags": ["facebook", "meta", "pytorch", "llama", "llama-3"], "model_name": "Llama 3 70B Instruct", "license_name": "llama3", "license_link": "LICENSE", "extra_gated_prompt": "### META LLAMA 3 COMMUNITY LICENSE AGREEMENT\nMeta Llama 3 Version Release Date: April 18, 2024\n\"Agreement\" means the terms and conditions for use, reproduction, distribution and modification of the Llama Materials set forth herein.\n\"Documentation\" means the specifications, manuals and documentation accompanying Meta Llama 3 distributed by Meta at https://llama.meta.com/get-started/.\n\"Licensee\" or \"you\" means you, or your employer or any other person or entity (if you are entering into this Agreement on such person or entity\u2019s behalf), of the age required under applicable laws, rules or regulations to provide legal consent and that has legal authority to bind your employer or such other person or entity if you are entering in this Agreement on their behalf.\n\"Meta Llama 3\" means the foundational large language models and software and algorithms, including machine-learning model code, trained model weights, inference-enabling code, training-enabling code, fine-tuning enabling code and other elements of the foregoing distributed by Meta at https://llama.meta.com/llama-downloads.\n\"Llama Materials\" means, collectively, Meta\u2019s proprietary Meta Llama 3 and Documentation (and any portion thereof) made available under this Agreement.\n\"Meta\" or \"we\" means Meta Platforms Ireland Limited (if you are located in or, if you are an entity, your principal place of business is in the EEA or Switzerland) and Meta Platforms, Inc. (if you are located outside of the EEA or Switzerland).\n \n1. License Rights and Redistribution.\na. Grant of Rights. You are granted a non-exclusive, worldwide, non-transferable and royalty-free limited license under Meta\u2019s intellectual property or other rights owned by Meta embodied in the Llama Materials to use, reproduce, distribute, copy, create derivative works of, and make modifications to the Llama Materials.\nb. Redistribution and Use.\ni. If you distribute or make available the Llama Materials (or any derivative works thereof), or a product or service that uses any of them, including another AI model, you shall (A) provide a copy of this Agreement with any such Llama Materials; and (B) prominently display \u201cBuilt with Meta Llama 3\u201d on a related website, user interface, blogpost, about page, or product documentation. If you use the Llama Materials to create, train, fine tune, or otherwise improve an AI model, which is distributed or made available, you shall also include \u201cLlama 3\u201d at the beginning of any such AI model name.\nii. If you receive Llama Materials, or any derivative works thereof, from a Licensee as part of an integrated end user product, then Section 2 of this Agreement will not apply to you.\niii. You must retain in all copies of the Llama Materials that you distribute the following attribution notice within a \u201cNotice\u201d text file distributed as a part of such copies: \u201cMeta Llama 3 is licensed under the Meta Llama 3 Community License, Copyright \u00a9 Meta Platforms, Inc. All Rights Reserved.\u201d\niv. Your use of the Llama Materials must comply with applicable laws and regulations (including trade compliance laws and regulations) and adhere to the Acceptable Use Policy for the Llama Materials (available at https://llama.meta.com/llama3/use-policy), which is hereby incorporated by reference into this Agreement.\nv. You will not use the Llama Materials or any output or results of the Llama Materials to improve any other large language model (excluding Meta Llama 3 or derivative works thereof).\n2. Additional Commercial Terms. If, on the Meta Llama 3 version release date, the monthly active users of the products or services made available by or for Licensee, or Licensee\u2019s affiliates, is greater than 700 million monthly active users in the preceding calendar month, you must request a license from Meta, which Meta may grant to you in its sole discretion, and you are not authorized to exercise any of the rights under this Agreement unless or until Meta otherwise expressly grants you such rights.\n3. Disclaimer of Warranty. UNLESS REQUIRED BY APPLICABLE LAW, THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS THEREFROM ARE PROVIDED ON AN \u201cAS IS\u201d BASIS, WITHOUT WARRANTIES OF ANY KIND, AND META DISCLAIMS ALL WARRANTIES OF ANY KIND, BOTH EXPRESS AND IMPLIED, INCLUDING, WITHOUT LIMITATION, ANY WARRANTIES OF TITLE, NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE. YOU ARE SOLELY RESPONSIBLE FOR DETERMINING THE APPROPRIATENESS OF USING OR REDISTRIBUTING THE LLAMA MATERIALS AND ASSUME ANY RISKS ASSOCIATED WITH YOUR USE OF THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS.\n4. Limitation of Liability. IN NO EVENT WILL META OR ITS AFFILIATES BE LIABLE UNDER ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, TORT, NEGLIGENCE, PRODUCTS LIABILITY, OR OTHERWISE, ARISING OUT OF THIS AGREEMENT, FOR ANY LOST PROFITS OR ANY INDIRECT, SPECIAL, CONSEQUENTIAL, INCIDENTAL, EXEMPLARY OR PUNITIVE DAMAGES, EVEN IF META OR ITS AFFILIATES HAVE BEEN ADVISED OF THE POSSIBILITY OF ANY OF THE FOREGOING.\n5. Intellectual Property.\na. No trademark licenses are granted under this Agreement, and in connection with the Llama Materials, neither Meta nor Licensee may use any name or mark owned by or associated with the other or any of its affiliates, except as required for reasonable and customary use in describing and redistributing the Llama Materials or as set forth in this Section 5(a). Meta hereby grants you a license to use \u201cLlama 3\u201d (the \u201cMark\u201d) solely as required to comply with the last sentence of Section 1.b.i. You will comply with Meta\u2019s brand guidelines (currently accessible at https://about.meta.com/brand/resources/meta/company-brand/ ). All goodwill arising out of your use of the Mark will inure to the benefit of Meta.\nb. Subject to Meta\u2019s ownership of Llama Materials and derivatives made by or for Meta, with respect to any derivative works and modifications of the Llama Materials that are made by you, as between you and Meta, you are and will be the owner of such derivative works and modifications.\nc. If you institute litigation or other proceedings against Meta or any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Llama Materials or Meta Llama 3 outputs or results, or any portion of any of the foregoing, constitutes infringement of intellectual property or other rights owned or licensable by you, then any licenses granted to you under this Agreement shall terminate as of the date such litigation or claim is filed or instituted. You will indemnify and hold harmless Meta from and against any claim by any third party arising out of or related to your use or distribution of the Llama Materials.\n6. Term and Termination. The term of this Agreement will commence upon your acceptance of this Agreement or access to the Llama Materials and will continue in full force and effect until terminated in accordance with the terms and conditions herein. Meta may terminate this Agreement if you are in breach of any term or condition of this Agreement. Upon termination of this Agreement, you shall delete and cease use of the Llama Materials. Sections 3, 4 and 7 shall survive the termination of this Agreement.\n7. Governing Law and Jurisdiction. This Agreement will be governed and construed under the laws of the State of California without regard to choice of law principles, and the UN Convention on Contracts for the International Sale of Goods does not apply to this Agreement. The courts of California shall have exclusive jurisdiction of any dispute arising out of this Agreement.\n### Meta Llama 3 Acceptable Use Policy\nMeta is committed to promoting safe and fair use of its tools and features, including Meta Llama 3. If you access or use Meta Llama 3, you agree to this Acceptable Use Policy (\u201cPolicy\u201d). The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy](https://llama.meta.com/llama3/use-policy)\n#### Prohibited Uses\nWe want everyone to use Meta Llama 3 safely and responsibly. You agree you will not use, or allow others to use, Meta Llama 3 to: 1. Violate the law or others\u2019 rights, including to:\n 1. Engage in, promote, generate, contribute to, encourage, plan, incite, or further illegal or unlawful activity or content, such as:\n 1. Violence or terrorism\n 2. Exploitation or harm to children, including the solicitation, creation, acquisition, or dissemination of child exploitative content or failure to report Child Sexual Abuse Material\n 3. Human trafficking, exploitation, and sexual violence\n 4. The illegal distribution of information or materials to minors, including obscene materials, or failure to employ legally required age-gating in connection with such information or materials.\n 5. Sexual solicitation\n 6. Any other criminal activity\n 2. Engage in, promote, incite, or facilitate the harassment, abuse, threatening, or bullying of individuals or groups of individuals\n 3. Engage in, promote, incite, or facilitate discrimination or other unlawful or harmful conduct in the provision of employment, employment benefits, credit, housing, other economic benefits, or other essential goods and services\n 4. Engage in the unauthorized or unlicensed practice of any profession including, but not limited to, financial, legal, medical/health, or related professional practices\n 5. Collect, process, disclose, generate, or infer health, demographic, or other sensitive personal or private information about individuals without rights and consents required by applicable laws\n 6. Engage in or facilitate any action or generate any content that infringes, misappropriates, or otherwise violates any third-party rights, including the outputs or results of any products or services using the Llama Materials\n 7. Create, generate, or facilitate the creation of malicious code, malware, computer viruses or do anything else that could disable, overburden, interfere with or impair the proper working, integrity, operation or appearance of a website or computer system\n2. Engage in, promote, incite, facilitate, or assist in the planning or development of activities that present a risk of death or bodily harm to individuals, including use of Meta Llama 3 related to the following:\n 1. Military, warfare, nuclear industries or applications, espionage, use for materials or activities that are subject to the International Traffic Arms Regulations (ITAR) maintained by the United States Department of State\n 2. Guns and illegal weapons (including weapon development)\n 3. Illegal drugs and regulated/controlled substances\n 4. Operation of critical infrastructure, transportation technologies, or heavy machinery\n 5. Self-harm or harm to others, including suicide, cutting, and eating disorders\n 6. Any content intended to incite or promote violence, abuse, or any infliction of bodily harm to an individual\n3. Intentionally deceive or mislead others, including use of Meta Llama 3 related to the following:\n 1. Generating, promoting, or furthering fraud or the creation or promotion of disinformation\n 2. Generating, promoting, or furthering defamatory content, including the creation of defamatory statements, images, or other content\n 3. Generating, promoting, or further distributing spam\n 4. Impersonating another individual without consent, authorization, or legal right\n 5. Representing that the use of Meta Llama 3 or outputs are human-generated\n 6. Generating or facilitating false online engagement, including fake reviews and other means of fake online engagement\n4. Fail to appropriately disclose to end users any known dangers of your AI system\nPlease report any violation of this Policy, software \u201cbug,\u201d or other problems that could lead to a violation of this Policy through one of the following means:\n * Reporting issues with the model: [https://github.com/meta-llama/llama3](https://github.com/meta-llama/llama3)\n * Reporting risky content generated by the model:\n developers.facebook.com/llama_output_feedback\n * Reporting bugs and security concerns: facebook.com/whitehat/info\n * Reporting violations of the Acceptable Use Policy or unlicensed uses of Meta Llama 3: [email protected]", "extra_gated_fields": {"First Name": "text", "Last Name": "text", "Date of birth": "date_picker", "Country": "country", "Affiliation": "text", "geo": "ip_location", "By clicking Submit below I accept the terms of the license and acknowledge that the information I provide will be collected stored processed and shared in accordance with the Meta Privacy Policy": "checkbox"}, "extra_gated_description": "The information you provide will be collected, stored, processed and shared in accordance with the [Meta Privacy Policy](https://www.facebook.com/privacy/policy/).", "extra_gated_button_content": "Submit", "base_model": "meta-llama/Meta-Llama-3-70B-Instruct", "inference": false, "model_creator": "Meta Llama 3", "model_link": "https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct", "model_type": "llama", "pipeline_tag": "text-generation", "quantized_by": "FriendliAI"}	FriendliAI/Meta-Llama-3-70B-Instruct-fp8	null	[ "transformers", "safetensors", "llama", "text-generation", "facebook", "meta", "pytorch", "llama-3", "conversational", "en", "base_model:meta-llama/Meta-Llama-3-70B-Instruct", "license:other", "autotrain_compatible", "text-generation-inference", "8-bit", "region:us" ]	null	2024-04-19T04:52:52+00:00	[]	[ "en" ]	TAGS #transformers #safetensors #llama #text-generation #facebook #meta #pytorch #llama-3 #conversational #en #base_model-meta-llama/Meta-Llama-3-70B-Instruct #license-other #autotrain_compatible #text-generation-inference #8-bit #region-us	![Friendli Logo](https://i.URL width=) Llama 3 70B Instruct - FP8 ========================== * Model creator: Meta Llama 3 * Original model: Llama 3 70B Instruct Description ----------- This repo contains the Llama 3 70B Instruct model quantized to FP8 by FriendliAI, significantly enhancing its inference efficiency while maintaining high accuracy. Note that FP8 is only supported by NVIDIA Ada, Hopper, and Blackwell GPU architectures. Check out FriendliAI documentation for more details. License ------- Refer to the license of the original model card. Compatibility ------------- This model is compatible with Friendli Container. Prerequisites ------------- * Before you begin, make sure you have signed up for Friendli Suite. You can use Friendli Containers free of charge for four weeks. * Prepare a Personal Access Token following this guide. * Prepare a Friendli Container Secret following this guide. ### Preparing Personal Access Token PAT (Personal Access Token) is the user credential for for logging into our container registry. 1. Sign in Friendli Suite. 2. Go to User Settings > Tokens and click 'Create new token'. 3. Save your created token value. ### Preparing Container Secret Container secret is a credential to launch our Friendli Container images. You should pass the container secret as an environment variable to run the container image. 1. Sign in Friendli Suite. 2. Go to Container > Container Secrets and click 'Create secret'. 3. Save your created secret value. ### Pulling Friendli Container Image 1. Log in to the Docker client using the personal access token created as outlined in this guide. 2. Pull image Running Friendli Container -------------------------- Once you've prepared the image of Friendli Container, you can launch it to create a serving endpoint. --- Original model card: Meta Llama 3 70B Instruct ============================================== Model Details ------------- Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety. Model developers Meta Variations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants. Input Models input text only. Output Models generate text and code only. Model Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. Llama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability. Model Release Date April 18, 2024. Status This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback. License A custom commercial license is available at: URL Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model README. For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go here. Intended Use ------------ Intended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. Out-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English. Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. How to use ---------- This repository contains two versions of Meta-Llama-3-70B-Instruct, for use with transformers and with the original 'llama3' codebase. ### Use with transformers See the snippet below for usage with Transformers: ### Use with 'llama3' Please, follow the instructions in the repository. To download Original checkpoints, see the example command below leveraging 'huggingface-cli': For Hugging Face support, we recommend using transformers or TGI, but a similar command works. Hardware and Software --------------------- Training Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute. Carbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program. CO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others. Training Data ------------- Overview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data. Data Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively. Benchmarks ---------- In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here. ### Base pretrained models ### Instruction tuned models ### Responsibility & Safety We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community. Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications. Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience. As part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started. #### Llama 3-Instruct As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case. Safety For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable. Refusals In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2. We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date. #### Responsible release In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision. Misuse If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL #### Critical risks CBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives) We have conducted a two fold assessment of the safety of the model in this area: * Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks. * Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model). ### Cyber Security We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability. ### Child Safety Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences. ### Community Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository. Finally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community. Ethical Considerations and Limitations -------------------------------------- The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress. But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety. Please see the Responsible Use Guide available at URL instructions @article{llama3modelcard, title={Llama 3 Model Card}, author={AI@Meta}, year={2024}, url = {URL } Contributors ------------ Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos	[ "### Preparing Personal Access Token\n\n\nPAT (Personal Access Token) is the user credential for for logging into our container registry.\n\n\n1. Sign in Friendli Suite.\n2. Go to User Settings > Tokens and click 'Create new token'.\n3. Save your created token value.", "### Preparing Container Secret\n\n\nContainer secret is a credential to launch our Friendli Container images.\nYou should pass the container secret as an environment variable to run the container image.\n\n\n1. Sign in Friendli Suite.\n2. Go to Container > Container Secrets and click 'Create secret'.\n3. Save your created secret value.", "### Pulling Friendli Container Image\n\n\n1. Log in to the Docker client using the personal access token created as outlined in this guide.\n2. Pull image\n\n\nRunning Friendli Container\n--------------------------\n\n\nOnce you've prepared the image of Friendli Container, you can launch it to create a serving endpoint.\n\n\n\n\n---\n\n\nOriginal model card: Meta Llama 3 70B Instruct\n==============================================\n\n\nModel Details\n-------------\n\n\nMeta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety.\n\n\nModel developers Meta\n\n\nVariations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants.\n\n\nInput Models input text only.\n\n\nOutput Models generate text and code only.\n\n\nModel Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.\n\n\n\nLlama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability.\n\n\nModel Release Date April 18, 2024.\n\n\nStatus This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback.\n\n\nLicense A custom commercial license is available at: URL\n\n\nWhere to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model README. For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go here.\n\n\nIntended Use\n------------\n\n\nIntended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks.\n\n\nOut-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English.\n\n\nNote: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy.\n\n\nHow to use\n----------\n\n\nThis repository contains two versions of Meta-Llama-3-70B-Instruct, for use with transformers and with the original 'llama3' codebase.", "### Use with transformers\n\n\nSee the snippet below for usage with Transformers:", "### Use with 'llama3'\n\n\nPlease, follow the instructions in the repository.\n\n\nTo download Original checkpoints, see the example command below leveraging 'huggingface-cli':\n\n\nFor Hugging Face support, we recommend using transformers or TGI, but a similar command works.\n\n\nHardware and Software\n---------------------\n\n\nTraining Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.\n\n\nCarbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.\n\n\n\nCO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.\n\n\nTraining Data\n-------------\n\n\nOverview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.\n\n\nData Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.\n\n\nBenchmarks\n----------\n\n\nIn this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here.", "### Base pretrained models", "### Instruction tuned models", "### Responsibility & Safety\n\n\nWe believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.\n\n\nFoundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.\n\n\nRather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.\n\n\nAs part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started.", "#### Llama 3-Instruct\n\n\nAs outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.\n\n\nSafety\n\n\nFor our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.\n\n\nRefusals\n\n\nIn addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.\n\n\nWe built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.", "#### Responsible release\n\n\nIn addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.\n\n\nMisuse\n\n\nIf you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL", "#### Critical risks\n\n\nCBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)\n\n\nWe have conducted a two fold assessment of the safety of the model in this area:\n\n\n* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.\n* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).", "### Cyber Security\n\n\nWe have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability.", "### Child Safety\n\n\nChild Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.", "### Community\n\n\nGenerative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository.\n\n\nFinally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community.\n\n\nEthical Considerations and Limitations\n--------------------------------------\n\n\nThe core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.\n\n\nBut Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.\n\n\nPlease see the Responsible Use Guide available at URL\n\n\ninstructions\n\n\n@article{llama3modelcard,\n\n\ntitle={Llama 3 Model Card},\n\n\nauthor={AI@Meta},\n\n\nyear={2024},\n\n\nurl = {URL\n\n\n}\n\n\nContributors\n------------\n\n\nAaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos" ]	[ "TAGS\n#transformers #safetensors #llama #text-generation #facebook #meta #pytorch #llama-3 #conversational #en #base_model-meta-llama/Meta-Llama-3-70B-Instruct #license-other #autotrain_compatible #text-generation-inference #8-bit #region-us \n", "### Preparing Personal Access Token\n\n\nPAT (Personal Access Token) is the user credential for for logging into our container registry.\n\n\n1. Sign in Friendli Suite.\n2. Go to User Settings > Tokens and click 'Create new token'.\n3. Save your created token value.", "### Preparing Container Secret\n\n\nContainer secret is a credential to launch our Friendli Container images.\nYou should pass the container secret as an environment variable to run the container image.\n\n\n1. Sign in Friendli Suite.\n2. Go to Container > Container Secrets and click 'Create secret'.\n3. Save your created secret value.", "### Pulling Friendli Container Image\n\n\n1. Log in to the Docker client using the personal access token created as outlined in this guide.\n2. Pull image\n\n\nRunning Friendli Container\n--------------------------\n\n\nOnce you've prepared the image of Friendli Container, you can launch it to create a serving endpoint.\n\n\n\n\n---\n\n\nOriginal model card: Meta Llama 3 70B Instruct\n==============================================\n\n\nModel Details\n-------------\n\n\nMeta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety.\n\n\nModel developers Meta\n\n\nVariations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants.\n\n\nInput Models input text only.\n\n\nOutput Models generate text and code only.\n\n\nModel Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.\n\n\n\nLlama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability.\n\n\nModel Release Date April 18, 2024.\n\n\nStatus This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback.\n\n\nLicense A custom commercial license is available at: URL\n\n\nWhere to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model README. For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go here.\n\n\nIntended Use\n------------\n\n\nIntended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks.\n\n\nOut-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English.\n\n\nNote: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy.\n\n\nHow to use\n----------\n\n\nThis repository contains two versions of Meta-Llama-3-70B-Instruct, for use with transformers and with the original 'llama3' codebase.", "### Use with transformers\n\n\nSee the snippet below for usage with Transformers:", "### Use with 'llama3'\n\n\nPlease, follow the instructions in the repository.\n\n\nTo download Original checkpoints, see the example command below leveraging 'huggingface-cli':\n\n\nFor Hugging Face support, we recommend using transformers or TGI, but a similar command works.\n\n\nHardware and Software\n---------------------\n\n\nTraining Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.\n\n\nCarbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.\n\n\n\nCO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.\n\n\nTraining Data\n-------------\n\n\nOverview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.\n\n\nData Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.\n\n\nBenchmarks\n----------\n\n\nIn this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here.", "### Base pretrained models", "### Instruction tuned models", "### Responsibility & Safety\n\n\nWe believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.\n\n\nFoundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.\n\n\nRather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.\n\n\nAs part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started.", "#### Llama 3-Instruct\n\n\nAs outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.\n\n\nSafety\n\n\nFor our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.\n\n\nRefusals\n\n\nIn addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.\n\n\nWe built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.", "#### Responsible release\n\n\nIn addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.\n\n\nMisuse\n\n\nIf you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL", "#### Critical risks\n\n\nCBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)\n\n\nWe have conducted a two fold assessment of the safety of the model in this area:\n\n\n* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.\n* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).", "### Cyber Security\n\n\nWe have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability.", "### Child Safety\n\n\nChild Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.", "### Community\n\n\nGenerative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository.\n\n\nFinally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community.\n\n\nEthical Considerations and Limitations\n--------------------------------------\n\n\nThe core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.\n\n\nBut Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.\n\n\nPlease see the Responsible Use Guide available at URL\n\n\ninstructions\n\n\n@article{llama3modelcard,\n\n\ntitle={Llama 3 Model Card},\n\n\nauthor={AI@Meta},\n\n\nyear={2024},\n\n\nurl = {URL\n\n\n}\n\n\nContributors\n------------\n\n\nAaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos" ]
reinforcement-learning	ml-agents	# ppo Agent playing SnowballTarget This is a trained model of a ppo agent playing SnowballTarget using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents). ## Usage (with ML-Agents) The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/ We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub: - A short tutorial where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction - A longer tutorial to understand how works ML-Agents: https://huggingface.co/learn/deep-rl-course/unit5/introduction ### Resume the training ```bash mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume ``` ### Watch your Agent play You can watch your agent playing directly in your browser 1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity 2. Step 1: Find your model_id: marcus07/ppo-SnowballTarget 3. Step 2: Select your .nn /.onnx file 4. Click on Watch the agent play 👀	{"library_name": "ml-agents", "tags": ["SnowballTarget", "deep-reinforcement-learning", "reinforcement-learning", "ML-Agents-SnowballTarget"]}	marcus07/ppo-SnowballTarget	null	[ "ml-agents", "tensorboard", "onnx", "SnowballTarget", "deep-reinforcement-learning", "reinforcement-learning", "ML-Agents-SnowballTarget", "region:us" ]	null	2024-04-19T04:53:08+00:00	[]	[]	TAGS #ml-agents #tensorboard #onnx #SnowballTarget #deep-reinforcement-learning #reinforcement-learning #ML-Agents-SnowballTarget #region-us	# ppo Agent playing SnowballTarget This is a trained model of a ppo agent playing SnowballTarget using the Unity ML-Agents Library. ## Usage (with ML-Agents) The Documentation: URL We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub: - A short tutorial where you teach Huggy the Dog to fetch the stick and then play with him directly in your browser: URL - A longer tutorial to understand how works ML-Agents: URL ### Resume the training ### Watch your Agent play You can watch your agent playing directly in your browser 1. If the environment is part of ML-Agents official environments, go to URL 2. Step 1: Find your model_id: marcus07/ppo-SnowballTarget 3. Step 2: Select your .nn /.onnx file 4. Click on Watch the agent play	[ "# ppo Agent playing SnowballTarget\n This is a trained model of a ppo agent playing SnowballTarget\n using the Unity ML-Agents Library.\n\n ## Usage (with ML-Agents)\n The Documentation: URL\n\n We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:\n - A short tutorial where you teach Huggy the Dog to fetch the stick and then play with him directly in your\n browser: URL\n - A longer tutorial to understand how works ML-Agents:\n URL\n\n ### Resume the training\n \n\n ### Watch your Agent play\n You can watch your agent playing directly in your browser\n\n 1. If the environment is part of ML-Agents official environments, go to URL\n 2. Step 1: Find your model_id: marcus07/ppo-SnowballTarget\n 3. Step 2: Select your .nn /.onnx file\n 4. Click on Watch the agent play" ]	[ "TAGS\n#ml-agents #tensorboard #onnx #SnowballTarget #deep-reinforcement-learning #reinforcement-learning #ML-Agents-SnowballTarget #region-us \n", "# ppo Agent playing SnowballTarget\n This is a trained model of a ppo agent playing SnowballTarget\n using the Unity ML-Agents Library.\n\n ## Usage (with ML-Agents)\n The Documentation: URL\n\n We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:\n - A short tutorial where you teach Huggy the Dog to fetch the stick and then play with him directly in your\n browser: URL\n - A longer tutorial to understand how works ML-Agents:\n URL\n\n ### Resume the training\n \n\n ### Watch your Agent play\n You can watch your agent playing directly in your browser\n\n 1. If the environment is part of ML-Agents official environments, go to URL\n 2. Step 1: Find your model_id: marcus07/ppo-SnowballTarget\n 3. Step 2: Select your .nn /.onnx file\n 4. Click on Watch the agent play" ]
text-generation	transformers	## Model Details Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety. Model developers Meta Variations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants. Input Models input text only. Output Models generate text and code only. Model Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. <table> <tr> <td> </td> <td><strong>Training Data</strong> </td> <td><strong>Params</strong> </td> <td><strong>Context length</strong> </td> <td><strong>GQA</strong> </td> <td><strong>Token count</strong> </td> <td><strong>Knowledge cutoff</strong> </td> </tr> <tr> <td rowspan="2" >Llama 3 </td> <td rowspan="2" >A new mix of publicly available online data. </td> <td>8B </td> <td>8k </td> <td>Yes </td> <td rowspan="2" >15T+ </td> <td>March, 2023 </td> </tr> <tr> <td>70B </td> <td>8k </td> <td>Yes </td> <td>December, 2023 </td> </tr> </table> Llama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability. Model Release Date April 18, 2024. Status This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback. License A custom commercial license is available at: [https://llama.meta.com/llama3/license](https://llama.meta.com/llama3/license) Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model [README](https://github.com/meta-llama/llama3). For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go [here](https://github.com/meta-llama/llama-recipes). ## Intended Use Intended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. Out-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English. Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. ## How to use This repository contains two versions of Meta-Llama-3-8B, for use with transformers and with the original `llama3` codebase. ### Use with transformers See the snippet below for usage with Transformers: ```python >>> import transformers >>> import torch >>> model_id = "meta-llama/Meta-Llama-3-8B" >>> pipeline = transformers.pipeline( "text-generation", model=model_id, model_kwargs={"torch_dtype": torch.bfloat16}, device_map="auto" ) >>> pipeline("Hey how are you doing today?") ``` ### Use with `llama3` Please, follow the instructions in the [repository](https://github.com/meta-llama/llama3). To download Original checkpoints, see the example command below leveraging `huggingface-cli`: ``` huggingface-cli download meta-llama/Meta-Llama-3-8B --include "original/" --local-dir Meta-Llama-3-8B ``` For Hugging Face support, we recommend using transformers or TGI, but a similar command works. ## Hardware and Software Training Factors* We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute. Carbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program. <table> <tr> <td> </td> <td><strong>Time (GPU hours)</strong> </td> <td><strong>Power Consumption (W)</strong> </td> <td><strong>Carbon Emitted(tCO2eq)</strong> </td> </tr> <tr> <td>Llama 3 8B </td> <td>1.3M </td> <td>700 </td> <td>390 </td> </tr> <tr> <td>Llama 3 70B </td> <td>6.4M </td> <td>700 </td> <td>1900 </td> </tr> <tr> <td>Total </td> <td>7.7M </td> <td> </td> <td>2290 </td> </tr> </table> CO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others. ## Training Data Overview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data. Data Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively. ## Benchmarks In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see [here](https://github.com/meta-llama/llama3/blob/main/eval_methodology.md). ### Base pretrained models <table> <tr> <td><strong>Category</strong> </td> <td><strong>Benchmark</strong> </td> <td><strong>Llama 3 8B</strong> </td> <td><strong>Llama2 7B</strong> </td> <td><strong>Llama2 13B</strong> </td> <td><strong>Llama 3 70B</strong> </td> <td><strong>Llama2 70B</strong> </td> </tr> <tr> <td rowspan="6" >General </td> <td>MMLU (5-shot) </td> <td>66.6 </td> <td>45.7 </td> <td>53.8 </td> <td>79.5 </td> <td>69.7 </td> </tr> <tr> <td>AGIEval English (3-5 shot) </td> <td>45.9 </td> <td>28.8 </td> <td>38.7 </td> <td>63.0 </td> <td>54.8 </td> </tr> <tr> <td>CommonSenseQA (7-shot) </td> <td>72.6 </td> <td>57.6 </td> <td>67.6 </td> <td>83.8 </td> <td>78.7 </td> </tr> <tr> <td>Winogrande (5-shot) </td> <td>76.1 </td> <td>73.3 </td> <td>75.4 </td> <td>83.1 </td> <td>81.8 </td> </tr> <tr> <td>BIG-Bench Hard (3-shot, CoT) </td> <td>61.1 </td> <td>38.1 </td> <td>47.0 </td> <td>81.3 </td> <td>65.7 </td> </tr> <tr> <td>ARC-Challenge (25-shot) </td> <td>78.6 </td> <td>53.7 </td> <td>67.6 </td> <td>93.0 </td> <td>85.3 </td> </tr> <tr> <td>Knowledge reasoning </td> <td>TriviaQA-Wiki (5-shot) </td> <td>78.5 </td> <td>72.1 </td> <td>79.6 </td> <td>89.7 </td> <td>87.5 </td> </tr> <tr> <td rowspan="4" >Reading comprehension </td> <td>SQuAD (1-shot) </td> <td>76.4 </td> <td>72.2 </td> <td>72.1 </td> <td>85.6 </td> <td>82.6 </td> </tr> <tr> <td>QuAC (1-shot, F1) </td> <td>44.4 </td> <td>39.6 </td> <td>44.9 </td> <td>51.1 </td> <td>49.4 </td> </tr> <tr> <td>BoolQ (0-shot) </td> <td>75.7 </td> <td>65.5 </td> <td>66.9 </td> <td>79.0 </td> <td>73.1 </td> </tr> <tr> <td>DROP (3-shot, F1) </td> <td>58.4 </td> <td>37.9 </td> <td>49.8 </td> <td>79.7 </td> <td>70.2 </td> </tr> </table> ### Instruction tuned models <table> <tr> <td><strong>Benchmark</strong> </td> <td><strong>Llama 3 8B</strong> </td> <td><strong>Llama 2 7B</strong> </td> <td><strong>Llama 2 13B</strong> </td> <td><strong>Llama 3 70B</strong> </td> <td><strong>Llama 2 70B</strong> </td> </tr> <tr> <td>MMLU (5-shot) </td> <td>68.4 </td> <td>34.1 </td> <td>47.8 </td> <td>82.0 </td> <td>52.9 </td> </tr> <tr> <td>GPQA (0-shot) </td> <td>34.2 </td> <td>21.7 </td> <td>22.3 </td> <td>39.5 </td> <td>21.0 </td> </tr> <tr> <td>HumanEval (0-shot) </td> <td>62.2 </td> <td>7.9 </td> <td>14.0 </td> <td>81.7 </td> <td>25.6 </td> </tr> <tr> <td>GSM-8K (8-shot, CoT) </td> <td>79.6 </td> <td>25.7 </td> <td>77.4 </td> <td>93.0 </td> <td>57.5 </td> </tr> <tr> <td>MATH (4-shot, CoT) </td> <td>30.0 </td> <td>3.8 </td> <td>6.7 </td> <td>50.4 </td> <td>11.6 </td> </tr> </table> ### Responsibility & Safety We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community. Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications. Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience. As part of the Llama 3 release, we updated our [Responsible Use Guide](https://llama.meta.com/responsible-use-guide/) to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including [Meta Llama Guard 2](https://llama.meta.com/purple-llama/) and [Code Shield](https://llama.meta.com/purple-llama/) safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a [reference implementation](https://github.com/meta-llama/llama-recipes/tree/main/recipes/responsible_ai) to get you started. #### Llama 3-Instruct As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case. <span style="text-decoration:underline;">Safety</span> For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable. <span style="text-decoration:underline;">Refusals</span> In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2. We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date. #### Responsible release In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision. Misuse If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy/](https://llama.meta.com/llama3/use-policy/). #### Critical risks <span style="text-decoration:underline;">CBRNE</span> (Chemical, Biological, Radiological, Nuclear, and high yield Explosives) We have conducted a two fold assessment of the safety of the model in this area: * Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks. * Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model). ### <span style="text-decoration:underline;">Cyber Security </span> We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of [equivalent coding capability](https://huggingface.co/spaces/facebook/CyberSecEval). ### <span style="text-decoration:underline;">Child Safety</span> Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences. ### Community Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our [Github repository](https://github.com/meta-llama/PurpleLlama). Finally, we put in place a set of resources including an [output reporting mechanism](https://developers.facebook.com/llama_output_feedback) and [bug bounty program](https://www.facebook.com/whitehat) to continuously improve the Llama technology with the help of the community. ## Ethical Considerations and Limitations The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress. But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating [Purple Llama](https://github.com/facebookresearch/PurpleLlama) solutions into your workflows and specifically [Llama Guard](https://ai.meta.com/research/publications/llama-guard-llm-based-input-output-safeguard-for-human-ai-conversations/) which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety. Please see the Responsible Use Guide available at [http://llama.meta.com/responsible-use-guide](http://llama.meta.com/responsible-use-guide) ## Citation instructions @article{llama3modelcard, title={Llama 3 Model Card}, author={AI@Meta}, year={2024}, url = {https://github.com/meta-llama/llama3/blob/main/MODEL_CARD.md} } ## Contributors Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos	{"language": ["en"], "license": "other", "tags": ["facebook", "meta", "pytorch", "llama", "llama-3"], "pipeline_tag": "text-generation", "license_name": "llama3", "license_link": "LICENSE", "extra_gated_prompt": "### META LLAMA 3 COMMUNITY LICENSE AGREEMENT\nMeta Llama 3 Version Release Date: April 18, 2024\n\"Agreement\" means the terms and conditions for use, reproduction, distribution and modification of the Llama Materials set forth herein.\n\"Documentation\" means the specifications, manuals and documentation accompanying Meta Llama 3 distributed by Meta at https://llama.meta.com/get-started/.\n\"Licensee\" or \"you\" means you, or your employer or any other person or entity (if you are entering into this Agreement on such person or entity\u2019s behalf), of the age required under applicable laws, rules or regulations to provide legal consent and that has legal authority to bind your employer or such other person or entity if you are entering in this Agreement on their behalf.\n\"Meta Llama 3\" means the foundational large language models and software and algorithms, including machine-learning model code, trained model weights, inference-enabling code, training-enabling code, fine-tuning enabling code and other elements of the foregoing distributed by Meta at https://llama.meta.com/llama-downloads.\n\"Llama Materials\" means, collectively, Meta\u2019s proprietary Meta Llama 3 and Documentation (and any portion thereof) made available under this Agreement.\n\"Meta\" or \"we\" means Meta Platforms Ireland Limited (if you are located in or, if you are an entity, your principal place of business is in the EEA or Switzerland) and Meta Platforms, Inc. (if you are located outside of the EEA or Switzerland).\n \n1. License Rights and Redistribution.\na. Grant of Rights. You are granted a non-exclusive, worldwide, non-transferable and royalty-free limited license under Meta\u2019s intellectual property or other rights owned by Meta embodied in the Llama Materials to use, reproduce, distribute, copy, create derivative works of, and make modifications to the Llama Materials.\nb. Redistribution and Use.\ni. If you distribute or make available the Llama Materials (or any derivative works thereof), or a product or service that uses any of them, including another AI model, you shall (A) provide a copy of this Agreement with any such Llama Materials; and (B) prominently display \u201cBuilt with Meta Llama 3\u201d on a related website, user interface, blogpost, about page, or product documentation. If you use the Llama Materials to create, train, fine tune, or otherwise improve an AI model, which is distributed or made available, you shall also include \u201cLlama 3\u201d at the beginning of any such AI model name.\nii. If you receive Llama Materials, or any derivative works thereof, from a Licensee as part of an integrated end user product, then Section 2 of this Agreement will not apply to you.\niii. You must retain in all copies of the Llama Materials that you distribute the following attribution notice within a \u201cNotice\u201d text file distributed as a part of such copies: \u201cMeta Llama 3 is licensed under the Meta Llama 3 Community License, Copyright \u00a9 Meta Platforms, Inc. All Rights Reserved.\u201d\niv. Your use of the Llama Materials must comply with applicable laws and regulations (including trade compliance laws and regulations) and adhere to the Acceptable Use Policy for the Llama Materials (available at https://llama.meta.com/llama3/use-policy), which is hereby incorporated by reference into this Agreement.\nv. You will not use the Llama Materials or any output or results of the Llama Materials to improve any other large language model (excluding Meta Llama 3 or derivative works thereof).\n2. Additional Commercial Terms. If, on the Meta Llama 3 version release date, the monthly active users of the products or services made available by or for Licensee, or Licensee\u2019s affiliates, is greater than 700 million monthly active users in the preceding calendar month, you must request a license from Meta, which Meta may grant to you in its sole discretion, and you are not authorized to exercise any of the rights under this Agreement unless or until Meta otherwise expressly grants you such rights.\n3. Disclaimer of Warranty. UNLESS REQUIRED BY APPLICABLE LAW, THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS THEREFROM ARE PROVIDED ON AN \u201cAS IS\u201d BASIS, WITHOUT WARRANTIES OF ANY KIND, AND META DISCLAIMS ALL WARRANTIES OF ANY KIND, BOTH EXPRESS AND IMPLIED, INCLUDING, WITHOUT LIMITATION, ANY WARRANTIES OF TITLE, NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE. YOU ARE SOLELY RESPONSIBLE FOR DETERMINING THE APPROPRIATENESS OF USING OR REDISTRIBUTING THE LLAMA MATERIALS AND ASSUME ANY RISKS ASSOCIATED WITH YOUR USE OF THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS.\n4. Limitation of Liability. IN NO EVENT WILL META OR ITS AFFILIATES BE LIABLE UNDER ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, TORT, NEGLIGENCE, PRODUCTS LIABILITY, OR OTHERWISE, ARISING OUT OF THIS AGREEMENT, FOR ANY LOST PROFITS OR ANY INDIRECT, SPECIAL, CONSEQUENTIAL, INCIDENTAL, EXEMPLARY OR PUNITIVE DAMAGES, EVEN IF META OR ITS AFFILIATES HAVE BEEN ADVISED OF THE POSSIBILITY OF ANY OF THE FOREGOING.\n5. Intellectual Property.\na. No trademark licenses are granted under this Agreement, and in connection with the Llama Materials, neither Meta nor Licensee may use any name or mark owned by or associated with the other or any of its affiliates, except as required for reasonable and customary use in describing and redistributing the Llama Materials or as set forth in this Section 5(a). Meta hereby grants you a license to use \u201cLlama 3\u201d (the \u201cMark\u201d) solely as required to comply with the last sentence of Section 1.b.i. You will comply with Meta\u2019s brand guidelines (currently accessible at https://about.meta.com/brand/resources/meta/company-brand/ ). All goodwill arising out of your use of the Mark will inure to the benefit of Meta.\nb. Subject to Meta\u2019s ownership of Llama Materials and derivatives made by or for Meta, with respect to any derivative works and modifications of the Llama Materials that are made by you, as between you and Meta, you are and will be the owner of such derivative works and modifications.\nc. If you institute litigation or other proceedings against Meta or any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Llama Materials or Meta Llama 3 outputs or results, or any portion of any of the foregoing, constitutes infringement of intellectual property or other rights owned or licensable by you, then any licenses granted to you under this Agreement shall terminate as of the date such litigation or claim is filed or instituted. You will indemnify and hold harmless Meta from and against any claim by any third party arising out of or related to your use or distribution of the Llama Materials.\n6. Term and Termination. The term of this Agreement will commence upon your acceptance of this Agreement or access to the Llama Materials and will continue in full force and effect until terminated in accordance with the terms and conditions herein. Meta may terminate this Agreement if you are in breach of any term or condition of this Agreement. Upon termination of this Agreement, you shall delete and cease use of the Llama Materials. Sections 3, 4 and 7 shall survive the termination of this Agreement.\n7. Governing Law and Jurisdiction. This Agreement will be governed and construed under the laws of the State of California without regard to choice of law principles, and the UN Convention on Contracts for the International Sale of Goods does not apply to this Agreement. The courts of California shall have exclusive jurisdiction of any dispute arising out of this Agreement.\n### Meta Llama 3 Acceptable Use Policy\nMeta is committed to promoting safe and fair use of its tools and features, including Meta Llama 3. If you access or use Meta Llama 3, you agree to this Acceptable Use Policy (\u201cPolicy\u201d). The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy](https://llama.meta.com/llama3/use-policy)\n#### Prohibited Uses\nWe want everyone to use Meta Llama 3 safely and responsibly. You agree you will not use, or allow others to use, Meta Llama 3 to: 1. Violate the law or others\u2019 rights, including to:\n 1. Engage in, promote, generate, contribute to, encourage, plan, incite, or further illegal or unlawful activity or content, such as:\n 1. Violence or terrorism\n 2. Exploitation or harm to children, including the solicitation, creation, acquisition, or dissemination of child exploitative content or failure to report Child Sexual Abuse Material\n 3. Human trafficking, exploitation, and sexual violence\n 4. The illegal distribution of information or materials to minors, including obscene materials, or failure to employ legally required age-gating in connection with such information or materials.\n 5. Sexual solicitation\n 6. Any other criminal activity\n 2. Engage in, promote, incite, or facilitate the harassment, abuse, threatening, or bullying of individuals or groups of individuals\n 3. Engage in, promote, incite, or facilitate discrimination or other unlawful or harmful conduct in the provision of employment, employment benefits, credit, housing, other economic benefits, or other essential goods and services\n 4. Engage in the unauthorized or unlicensed practice of any profession including, but not limited to, financial, legal, medical/health, or related professional practices\n 5. Collect, process, disclose, generate, or infer health, demographic, or other sensitive personal or private information about individuals without rights and consents required by applicable laws\n 6. Engage in or facilitate any action or generate any content that infringes, misappropriates, or otherwise violates any third-party rights, including the outputs or results of any products or services using the Llama Materials\n 7. Create, generate, or facilitate the creation of malicious code, malware, computer viruses or do anything else that could disable, overburden, interfere with or impair the proper working, integrity, operation or appearance of a website or computer system\n2. Engage in, promote, incite, facilitate, or assist in the planning or development of activities that present a risk of death or bodily harm to individuals, including use of Meta Llama 3 related to the following:\n 1. Military, warfare, nuclear industries or applications, espionage, use for materials or activities that are subject to the International Traffic Arms Regulations (ITAR) maintained by the United States Department of State\n 2. Guns and illegal weapons (including weapon development)\n 3. Illegal drugs and regulated/controlled substances\n 4. Operation of critical infrastructure, transportation technologies, or heavy machinery\n 5. Self-harm or harm to others, including suicide, cutting, and eating disorders\n 6. Any content intended to incite or promote violence, abuse, or any infliction of bodily harm to an individual\n3. Intentionally deceive or mislead others, including use of Meta Llama 3 related to the following:\n 1. Generating, promoting, or furthering fraud or the creation or promotion of disinformation\n 2. Generating, promoting, or furthering defamatory content, including the creation of defamatory statements, images, or other content\n 3. Generating, promoting, or further distributing spam\n 4. Impersonating another individual without consent, authorization, or legal right\n 5. Representing that the use of Meta Llama 3 or outputs are human-generated\n 6. Generating or facilitating false online engagement, including fake reviews and other means of fake online engagement\n4. Fail to appropriately disclose to end users any known dangers of your AI system\nPlease report any violation of this Policy, software \u201cbug,\u201d or other problems that could lead to a violation of this Policy through one of the following means:\n * Reporting issues with the model: [https://github.com/meta-llama/llama3](https://github.com/meta-llama/llama3)\n * Reporting risky content generated by the model:\n developers.facebook.com/llama_output_feedback\n * Reporting bugs and security concerns: facebook.com/whitehat/info\n * Reporting violations of the Acceptable Use Policy or unlicensed uses of Meta Llama 3: [email protected]", "extra_gated_fields": {"First Name": "text", "Last Name": "text", "Date of birth": "date_picker", "Country": "country", "Affiliation": "text", "geo": "ip_location", "By clicking Submit below I accept the terms of the license and acknowledge that the information I provide will be collected stored processed and shared in accordance with the Meta Privacy Policy": "checkbox"}, "extra_gated_description": "The information you provide will be collected, stored, processed and shared in accordance with the [Meta Privacy Policy](https://www.facebook.com/privacy/policy/).", "extra_gated_button_content": "Submit"}	axolotl-ai-co/llama-3-8b-chatml	null	[ "transformers", "safetensors", "llama", "text-generation", "facebook", "meta", "pytorch", "llama-3", "conversational", "en", "license:other", "autotrain_compatible", "endpoints_compatible", "text-generation-inference", "region:us" ]	null	2024-04-19T04:54:29+00:00	[]	[ "en" ]	TAGS #transformers #safetensors #llama #text-generation #facebook #meta #pytorch #llama-3 #conversational #en #license-other #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us	Model Details ------------- Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety. Model developers Meta Variations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants. Input Models input text only. Output Models generate text and code only. Model Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. Llama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability. Model Release Date April 18, 2024. Status This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback. License A custom commercial license is available at: URL Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model README. For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go here. Intended Use ------------ Intended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. Out-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English. Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. How to use ---------- This repository contains two versions of Meta-Llama-3-8B, for use with transformers and with the original 'llama3' codebase. ### Use with transformers See the snippet below for usage with Transformers: ### Use with 'llama3' Please, follow the instructions in the repository. To download Original checkpoints, see the example command below leveraging 'huggingface-cli': For Hugging Face support, we recommend using transformers or TGI, but a similar command works. Hardware and Software --------------------- Training Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute. Carbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program. CO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others. Training Data ------------- Overview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data. Data Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively. Benchmarks ---------- In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here. ### Base pretrained models ### Instruction tuned models ### Responsibility & Safety We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community. Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications. Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience. As part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started. #### Llama 3-Instruct As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case. Safety For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable. Refusals In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2. We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date. #### Responsible release In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision. Misuse If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL #### Critical risks CBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives) We have conducted a two fold assessment of the safety of the model in this area: * Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks. * Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model). ### Cyber Security We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability. ### Child Safety Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences. ### Community Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository. Finally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community. Ethical Considerations and Limitations -------------------------------------- The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress. But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety. Please see the Responsible Use Guide available at URL instructions @article{llama3modelcard, title={Llama 3 Model Card}, author={AI@Meta}, year={2024}, url = {URL } Contributors ------------ Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos	[ "### Use with transformers\n\n\nSee the snippet below for usage with Transformers:", "### Use with 'llama3'\n\n\nPlease, follow the instructions in the repository.\n\n\nTo download Original checkpoints, see the example command below leveraging 'huggingface-cli':\n\n\nFor Hugging Face support, we recommend using transformers or TGI, but a similar command works.\n\n\nHardware and Software\n---------------------\n\n\nTraining Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.\n\n\nCarbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.\n\n\n\nCO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.\n\n\nTraining Data\n-------------\n\n\nOverview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.\n\n\nData Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.\n\n\nBenchmarks\n----------\n\n\nIn this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here.", "### Base pretrained models", "### Instruction tuned models", "### Responsibility & Safety\n\n\nWe believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.\n\n\nFoundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.\n\n\nRather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.\n\n\nAs part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started.", "#### Llama 3-Instruct\n\n\nAs outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.\n\n\nSafety\n\n\nFor our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.\n\n\nRefusals\n\n\nIn addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.\n\n\nWe built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.", "#### Responsible release\n\n\nIn addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.\n\n\nMisuse\n\n\nIf you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL", "#### Critical risks\n\n\nCBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)\n\n\nWe have conducted a two fold assessment of the safety of the model in this area:\n\n\n* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.\n* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).", "### Cyber Security\n\n\nWe have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability.", "### Child Safety\n\n\nChild Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.", "### Community\n\n\nGenerative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository.\n\n\nFinally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community.\n\n\nEthical Considerations and Limitations\n--------------------------------------\n\n\nThe core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.\n\n\nBut Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.\n\n\nPlease see the Responsible Use Guide available at URL\n\n\ninstructions\n\n\n@article{llama3modelcard,\n\n\ntitle={Llama 3 Model Card},\n\n\nauthor={AI@Meta},\n\n\nyear={2024},\n\n\nurl = {URL\n\n\n}\n\n\nContributors\n------------\n\n\nAaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos" ]	[ "TAGS\n#transformers #safetensors #llama #text-generation #facebook #meta #pytorch #llama-3 #conversational #en #license-other #autotrain_compatible #endpoints_compatible #text-generation-inference #region-us \n", "### Use with transformers\n\n\nSee the snippet below for usage with Transformers:", "### Use with 'llama3'\n\n\nPlease, follow the instructions in the repository.\n\n\nTo download Original checkpoints, see the example command below leveraging 'huggingface-cli':\n\n\nFor Hugging Face support, we recommend using transformers or TGI, but a similar command works.\n\n\nHardware and Software\n---------------------\n\n\nTraining Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.\n\n\nCarbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.\n\n\n\nCO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.\n\n\nTraining Data\n-------------\n\n\nOverview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.\n\n\nData Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.\n\n\nBenchmarks\n----------\n\n\nIn this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here.", "### Base pretrained models", "### Instruction tuned models", "### Responsibility & Safety\n\n\nWe believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.\n\n\nFoundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.\n\n\nRather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.\n\n\nAs part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started.", "#### Llama 3-Instruct\n\n\nAs outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.\n\n\nSafety\n\n\nFor our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.\n\n\nRefusals\n\n\nIn addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.\n\n\nWe built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.", "#### Responsible release\n\n\nIn addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.\n\n\nMisuse\n\n\nIf you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL", "#### Critical risks\n\n\nCBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)\n\n\nWe have conducted a two fold assessment of the safety of the model in this area:\n\n\n* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.\n* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).", "### Cyber Security\n\n\nWe have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability.", "### Child Safety\n\n\nChild Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.", "### Community\n\n\nGenerative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository.\n\n\nFinally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community.\n\n\nEthical Considerations and Limitations\n--------------------------------------\n\n\nThe core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.\n\n\nBut Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.\n\n\nPlease see the Responsible Use Guide available at URL\n\n\ninstructions\n\n\n@article{llama3modelcard,\n\n\ntitle={Llama 3 Model Card},\n\n\nauthor={AI@Meta},\n\n\nyear={2024},\n\n\nurl = {URL\n\n\n}\n\n\nContributors\n------------\n\n\nAaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos" ]
text-generation	transformers	<!-- header start --> <p align="center"> <img src="https://i.imgur.com/mNM6Cai.png" width="100%" alt="Friendli Logo"> </p> <!-- header end --> # Llama 3 8B Instruct - FP8 - Model creator: [Meta Llama 3](https://huggingface.co/meta-llama) - Original model: [Llama 3 8B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) ## Description This repo contains the Llama 3 8B Instruct model quantized to FP8 by FriendliAI, significantly enhancing its inference efficiency while maintaining high accuracy. Note that FP8 is only supported by NVIDIA Ada, Hopper, and Blackwell GPU architectures. Check out [FriendliAI documentation](https://docs.friendli.ai/) for more details. ## License Refer to the license of the original model card. ## Compatibility This model is compatible with [Friendli Container](https://friendli.ai/products/container/). ## Prerequisites - Before you begin, make sure you have signed up for [Friendli Suite](https://suite.friendli.ai/). You can use Friendli Containers free of charge for four weeks. - Prepare a Personal Access Token following [this guide](#preparing-personal-access-token). - Prepare a Friendli Container Secret following [this guide](#preparing-container-secret). ### Preparing Personal Access Token PAT (Personal Access Token) is the user credential for for logging into our container registry. 1. Sign in [Friendli Suite](https://suite.friendli.ai/). 2. Go to [User Settings > Tokens](https://suite.friendli.ai/user-settings/tokens) and click 'Create new token'. 3. Save your created token value. ### Preparing Container Secret Container secret is a credential to launch our Friendli Container images. You should pass the container secret as an environment variable to run the container image. 1. Sign in [Friendli Suite](https://suite.friendli.ai/). 2. Go to Container > Container Secrets and click 'Create secret'. 3. Save your created secret value. ### Pulling Friendli Container Image 1. Log in to the Docker client using the personal access token created as outlined in [this guide](#preparing-personal-access-token). ```sh export FRIENDLI_PAT="YOUR PAT" docker login registry.friendli.ai -u $YOUR_EMAIL -p $FRIENDLI_PAT ``` 2. Pull image ```sh docker pull registry.friendli.ai/trial ``` ## Running Friendli Container Once you've prepared the image of Friendli Container, you can launch it to create a serving endpoint. ```sh docker run \ --gpus '"device=0"' \ -p 8000:8000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ -e FRIENDLI_CONTAINER_SECRET="YOUR CONTAINER SECRET" \ registry.friendli.ai/trial \ --web-server-port 8000 \ --hf-model-name FriendliAI/Meta-Llama-3-8B-Instruct-fp8 ``` --- # Original model card: Meta Llama 3 8B Instruct ## Model Details Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety. Model developers Meta Variations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants. Input Models input text only. Output Models generate text and code only. Model Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. <table> <tr> <td> </td> <td><strong>Training Data</strong> </td> <td><strong>Params</strong> </td> <td><strong>Context length</strong> </td> <td><strong>GQA</strong> </td> <td><strong>Token count</strong> </td> <td><strong>Knowledge cutoff</strong> </td> </tr> <tr> <td rowspan="2" >Llama 3 </td> <td rowspan="2" >A new mix of publicly available online data. </td> <td>8B </td> <td>8k </td> <td>Yes </td> <td rowspan="2" >15T+ </td> <td>March, 2023 </td> </tr> <tr> <td>70B </td> <td>8k </td> <td>Yes </td> <td>December, 2023 </td> </tr> </table> Llama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability. Model Release Date April 18, 2024. Status This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback. License A custom commercial license is available at: [https://llama.meta.com/llama3/license](https://llama.meta.com/llama3/license) Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model [README](https://github.com/meta-llama/llama3). For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go [here](https://github.com/meta-llama/llama-recipes). ## Intended Use Intended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. Out-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English. Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. ## How to use This repository contains two versions of Meta-Llama-3-8B-Instruct, for use with transformers and with the original `llama3` codebase. ### Use with transformers You can run conversational inference using the Transformers pipeline abstraction, or by leveraging the Auto classes with the `generate()` function. Let's see examples of both. #### Transformers pipeline ```python import transformers import torch model_id = "meta-llama/Meta-Llama-3-8B-Instruct" pipeline = transformers.pipeline( "text-generation", model=model_id, model_kwargs={"torch_dtype": torch.bfloat16}, device="auto", ) messages = [ {"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"}, {"role": "user", "content": "Who are you?"}, ] prompt = pipeline.tokenizer.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) terminators = [ pipeline.tokenizer.eos_token_id, pipeline.tokenizer.convert_tokens_to_ids("<\|eot_id\|>") ] outputs = pipeline( prompt, max_new_tokens=256, eos_token_id=terminators, do_sample=True, temperature=0.6, top_p=0.9, ) print(outputs[0]["generated_text"][len(prompt):]) ``` #### Transformers AutoModelForCausalLM ```python from transformers import AutoTokenizer, AutoModelForCausalLM import torch model_id = "meta-llama/Meta-Llama-3-8B-Instruct" tokenizer = AutoTokenizer.from_pretrained(model_id) model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.bfloat16, device_map="auto", ) messages = [ {"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"}, {"role": "user", "content": "Who are you?"}, ] input_ids = tokenizer.apply_chat_template( messages, add_generation_prompt=True, return_tensors="pt" ).to(model.device) terminators = [ tokenizer.eos_token_id, tokenizer.convert_tokens_to_ids("<\|eot_id\|>") ] outputs = model.generate( input_ids, max_new_tokens=256, eos_token_id=terminators, do_sample=True, temperature=0.6, top_p=0.9, ) response = outputs[0][input_ids.shape[-1]:] print(tokenizer.decode(response, skip_special_tokens=True)) ``` ### Use with `llama3` Please, follow the instructions in the [repository](https://github.com/meta-llama/llama3) To download Original checkpoints, see the example command below leveraging `huggingface-cli`: ``` huggingface-cli download meta-llama/Meta-Llama-3-8B-Instruct --include "original/" --local-dir Meta-Llama-3-8B-Instruct ``` For Hugging Face support, we recommend using transformers or TGI, but a similar command works. ## Hardware and Software Training Factors* We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute. Carbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program. <table> <tr> <td> </td> <td><strong>Time (GPU hours)</strong> </td> <td><strong>Power Consumption (W)</strong> </td> <td><strong>Carbon Emitted(tCO2eq)</strong> </td> </tr> <tr> <td>Llama 3 8B </td> <td>1.3M </td> <td>700 </td> <td>390 </td> </tr> <tr> <td>Llama 3 70B </td> <td>6.4M </td> <td>700 </td> <td>1900 </td> </tr> <tr> <td>Total </td> <td>7.7M </td> <td> </td> <td>2290 </td> </tr> </table> CO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others. ## Training Data Overview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data. Data Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively. ## Benchmarks In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see [here](https://github.com/meta-llama/llama3/blob/main/eval_methodology.md). ### Base pretrained models <table> <tr> <td><strong>Category</strong> </td> <td><strong>Benchmark</strong> </td> <td><strong>Llama 3 8B</strong> </td> <td><strong>Llama2 7B</strong> </td> <td><strong>Llama2 13B</strong> </td> <td><strong>Llama 3 70B</strong> </td> <td><strong>Llama2 70B</strong> </td> </tr> <tr> <td rowspan="6" >General </td> <td>MMLU (5-shot) </td> <td>66.6 </td> <td>45.7 </td> <td>53.8 </td> <td>79.5 </td> <td>69.7 </td> </tr> <tr> <td>AGIEval English (3-5 shot) </td> <td>45.9 </td> <td>28.8 </td> <td>38.7 </td> <td>63.0 </td> <td>54.8 </td> </tr> <tr> <td>CommonSenseQA (7-shot) </td> <td>72.6 </td> <td>57.6 </td> <td>67.6 </td> <td>83.8 </td> <td>78.7 </td> </tr> <tr> <td>Winogrande (5-shot) </td> <td>76.1 </td> <td>73.3 </td> <td>75.4 </td> <td>83.1 </td> <td>81.8 </td> </tr> <tr> <td>BIG-Bench Hard (3-shot, CoT) </td> <td>61.1 </td> <td>38.1 </td> <td>47.0 </td> <td>81.3 </td> <td>65.7 </td> </tr> <tr> <td>ARC-Challenge (25-shot) </td> <td>78.6 </td> <td>53.7 </td> <td>67.6 </td> <td>93.0 </td> <td>85.3 </td> </tr> <tr> <td>Knowledge reasoning </td> <td>TriviaQA-Wiki (5-shot) </td> <td>78.5 </td> <td>72.1 </td> <td>79.6 </td> <td>89.7 </td> <td>87.5 </td> </tr> <tr> <td rowspan="4" >Reading comprehension </td> <td>SQuAD (1-shot) </td> <td>76.4 </td> <td>72.2 </td> <td>72.1 </td> <td>85.6 </td> <td>82.6 </td> </tr> <tr> <td>QuAC (1-shot, F1) </td> <td>44.4 </td> <td>39.6 </td> <td>44.9 </td> <td>51.1 </td> <td>49.4 </td> </tr> <tr> <td>BoolQ (0-shot) </td> <td>75.7 </td> <td>65.5 </td> <td>66.9 </td> <td>79.0 </td> <td>73.1 </td> </tr> <tr> <td>DROP (3-shot, F1) </td> <td>58.4 </td> <td>37.9 </td> <td>49.8 </td> <td>79.7 </td> <td>70.2 </td> </tr> </table> ### Instruction tuned models <table> <tr> <td><strong>Benchmark</strong> </td> <td><strong>Llama 3 8B</strong> </td> <td><strong>Llama 2 7B</strong> </td> <td><strong>Llama 2 13B</strong> </td> <td><strong>Llama 3 70B</strong> </td> <td><strong>Llama 2 70B</strong> </td> </tr> <tr> <td>MMLU (5-shot) </td> <td>68.4 </td> <td>34.1 </td> <td>47.8 </td> <td>82.0 </td> <td>52.9 </td> </tr> <tr> <td>GPQA (0-shot) </td> <td>34.2 </td> <td>21.7 </td> <td>22.3 </td> <td>39.5 </td> <td>21.0 </td> </tr> <tr> <td>HumanEval (0-shot) </td> <td>62.2 </td> <td>7.9 </td> <td>14.0 </td> <td>81.7 </td> <td>25.6 </td> </tr> <tr> <td>GSM-8K (8-shot, CoT) </td> <td>79.6 </td> <td>25.7 </td> <td>77.4 </td> <td>93.0 </td> <td>57.5 </td> </tr> <tr> <td>MATH (4-shot, CoT) </td> <td>30.0 </td> <td>3.8 </td> <td>6.7 </td> <td>50.4 </td> <td>11.6 </td> </tr> </table> ### Responsibility & Safety We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community. Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications. Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience. As part of the Llama 3 release, we updated our [Responsible Use Guide](https://llama.meta.com/responsible-use-guide/) to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including [Meta Llama Guard 2](https://llama.meta.com/purple-llama/) and [Code Shield](https://llama.meta.com/purple-llama/) safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a [reference implementation](https://github.com/meta-llama/llama-recipes/tree/main/recipes/responsible_ai) to get you started. #### Llama 3-Instruct As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case. <span style="text-decoration:underline;">Safety</span> For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable. <span style="text-decoration:underline;">Refusals</span> In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2. We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date. #### Responsible release In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision. Misuse If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy/](https://llama.meta.com/llama3/use-policy/). #### Critical risks <span style="text-decoration:underline;">CBRNE</span> (Chemical, Biological, Radiological, Nuclear, and high yield Explosives) We have conducted a two fold assessment of the safety of the model in this area: * Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks. * Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model). ### <span style="text-decoration:underline;">Cyber Security </span> We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of [equivalent coding capability](https://huggingface.co/spaces/facebook/CyberSecEval). ### <span style="text-decoration:underline;">Child Safety</span> Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences. ### Community Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our [Github repository](https://github.com/meta-llama/PurpleLlama). Finally, we put in place a set of resources including an [output reporting mechanism](https://developers.facebook.com/llama_output_feedback) and [bug bounty program](https://www.facebook.com/whitehat) to continuously improve the Llama technology with the help of the community. ## Ethical Considerations and Limitations The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress. But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating [Purple Llama](https://github.com/facebookresearch/PurpleLlama) solutions into your workflows and specifically [Llama Guard](https://ai.meta.com/research/publications/llama-guard-llm-based-input-output-safeguard-for-human-ai-conversations/) which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety. Please see the Responsible Use Guide available at [http://llama.meta.com/responsible-use-guide](http://llama.meta.com/responsible-use-guide) ## Citation instructions @article{llama3modelcard, title={Llama 3 Model Card}, author={AI@Meta}, year={2024}, url = {https://github.com/meta-llama/llama3/blob/main/MODEL_CARD.md} } ## Contributors Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos	{"language": ["en"], "license": "other", "tags": ["facebook", "meta", "pytorch", "llama", "llama-3"], "model_name": "Llama 3 8B Instruct", "license_name": "llama3", "license_link": "LICENSE", "extra_gated_prompt": "### META LLAMA 3 COMMUNITY LICENSE AGREEMENT\nMeta Llama 3 Version Release Date: April 18, 2024\n\"Agreement\" means the terms and conditions for use, reproduction, distribution and modification of the Llama Materials set forth herein.\n\"Documentation\" means the specifications, manuals and documentation accompanying Meta Llama 3 distributed by Meta at https://llama.meta.com/get-started/.\n\"Licensee\" or \"you\" means you, or your employer or any other person or entity (if you are entering into this Agreement on such person or entity\u2019s behalf), of the age required under applicable laws, rules or regulations to provide legal consent and that has legal authority to bind your employer or such other person or entity if you are entering in this Agreement on their behalf.\n\"Meta Llama 3\" means the foundational large language models and software and algorithms, including machine-learning model code, trained model weights, inference-enabling code, training-enabling code, fine-tuning enabling code and other elements of the foregoing distributed by Meta at https://llama.meta.com/llama-downloads.\n\"Llama Materials\" means, collectively, Meta\u2019s proprietary Meta Llama 3 and Documentation (and any portion thereof) made available under this Agreement.\n\"Meta\" or \"we\" means Meta Platforms Ireland Limited (if you are located in or, if you are an entity, your principal place of business is in the EEA or Switzerland) and Meta Platforms, Inc. (if you are located outside of the EEA or Switzerland).\n \n1. License Rights and Redistribution.\na. Grant of Rights. You are granted a non-exclusive, worldwide, non-transferable and royalty-free limited license under Meta\u2019s intellectual property or other rights owned by Meta embodied in the Llama Materials to use, reproduce, distribute, copy, create derivative works of, and make modifications to the Llama Materials.\nb. Redistribution and Use.\ni. If you distribute or make available the Llama Materials (or any derivative works thereof), or a product or service that uses any of them, including another AI model, you shall (A) provide a copy of this Agreement with any such Llama Materials; and (B) prominently display \u201cBuilt with Meta Llama 3\u201d on a related website, user interface, blogpost, about page, or product documentation. If you use the Llama Materials to create, train, fine tune, or otherwise improve an AI model, which is distributed or made available, you shall also include \u201cLlama 3\u201d at the beginning of any such AI model name.\nii. If you receive Llama Materials, or any derivative works thereof, from a Licensee as part of an integrated end user product, then Section 2 of this Agreement will not apply to you.\niii. You must retain in all copies of the Llama Materials that you distribute the following attribution notice within a \u201cNotice\u201d text file distributed as a part of such copies: \u201cMeta Llama 3 is licensed under the Meta Llama 3 Community License, Copyright \u00a9 Meta Platforms, Inc. All Rights Reserved.\u201d\niv. Your use of the Llama Materials must comply with applicable laws and regulations (including trade compliance laws and regulations) and adhere to the Acceptable Use Policy for the Llama Materials (available at https://llama.meta.com/llama3/use-policy), which is hereby incorporated by reference into this Agreement.\nv. You will not use the Llama Materials or any output or results of the Llama Materials to improve any other large language model (excluding Meta Llama 3 or derivative works thereof).\n2. Additional Commercial Terms. If, on the Meta Llama 3 version release date, the monthly active users of the products or services made available by or for Licensee, or Licensee\u2019s affiliates, is greater than 700 million monthly active users in the preceding calendar month, you must request a license from Meta, which Meta may grant to you in its sole discretion, and you are not authorized to exercise any of the rights under this Agreement unless or until Meta otherwise expressly grants you such rights.\n3. Disclaimer of Warranty. UNLESS REQUIRED BY APPLICABLE LAW, THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS THEREFROM ARE PROVIDED ON AN \u201cAS IS\u201d BASIS, WITHOUT WARRANTIES OF ANY KIND, AND META DISCLAIMS ALL WARRANTIES OF ANY KIND, BOTH EXPRESS AND IMPLIED, INCLUDING, WITHOUT LIMITATION, ANY WARRANTIES OF TITLE, NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE. YOU ARE SOLELY RESPONSIBLE FOR DETERMINING THE APPROPRIATENESS OF USING OR REDISTRIBUTING THE LLAMA MATERIALS AND ASSUME ANY RISKS ASSOCIATED WITH YOUR USE OF THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS.\n4. Limitation of Liability. IN NO EVENT WILL META OR ITS AFFILIATES BE LIABLE UNDER ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, TORT, NEGLIGENCE, PRODUCTS LIABILITY, OR OTHERWISE, ARISING OUT OF THIS AGREEMENT, FOR ANY LOST PROFITS OR ANY INDIRECT, SPECIAL, CONSEQUENTIAL, INCIDENTAL, EXEMPLARY OR PUNITIVE DAMAGES, EVEN IF META OR ITS AFFILIATES HAVE BEEN ADVISED OF THE POSSIBILITY OF ANY OF THE FOREGOING.\n5. Intellectual Property.\na. No trademark licenses are granted under this Agreement, and in connection with the Llama Materials, neither Meta nor Licensee may use any name or mark owned by or associated with the other or any of its affiliates, except as required for reasonable and customary use in describing and redistributing the Llama Materials or as set forth in this Section 5(a). Meta hereby grants you a license to use \u201cLlama 3\u201d (the \u201cMark\u201d) solely as required to comply with the last sentence of Section 1.b.i. You will comply with Meta\u2019s brand guidelines (currently accessible at https://about.meta.com/brand/resources/meta/company-brand/ ). All goodwill arising out of your use of the Mark will inure to the benefit of Meta.\nb. Subject to Meta\u2019s ownership of Llama Materials and derivatives made by or for Meta, with respect to any derivative works and modifications of the Llama Materials that are made by you, as between you and Meta, you are and will be the owner of such derivative works and modifications.\nc. If you institute litigation or other proceedings against Meta or any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Llama Materials or Meta Llama 3 outputs or results, or any portion of any of the foregoing, constitutes infringement of intellectual property or other rights owned or licensable by you, then any licenses granted to you under this Agreement shall terminate as of the date such litigation or claim is filed or instituted. You will indemnify and hold harmless Meta from and against any claim by any third party arising out of or related to your use or distribution of the Llama Materials.\n6. Term and Termination. The term of this Agreement will commence upon your acceptance of this Agreement or access to the Llama Materials and will continue in full force and effect until terminated in accordance with the terms and conditions herein. Meta may terminate this Agreement if you are in breach of any term or condition of this Agreement. Upon termination of this Agreement, you shall delete and cease use of the Llama Materials. Sections 3, 4 and 7 shall survive the termination of this Agreement.\n7. Governing Law and Jurisdiction. This Agreement will be governed and construed under the laws of the State of California without regard to choice of law principles, and the UN Convention on Contracts for the International Sale of Goods does not apply to this Agreement. The courts of California shall have exclusive jurisdiction of any dispute arising out of this Agreement.\n### Meta Llama 3 Acceptable Use Policy\nMeta is committed to promoting safe and fair use of its tools and features, including Meta Llama 3. If you access or use Meta Llama 3, you agree to this Acceptable Use Policy (\u201cPolicy\u201d). The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy](https://llama.meta.com/llama3/use-policy)\n#### Prohibited Uses\nWe want everyone to use Meta Llama 3 safely and responsibly. You agree you will not use, or allow others to use, Meta Llama 3 to: 1. Violate the law or others\u2019 rights, including to:\n 1. Engage in, promote, generate, contribute to, encourage, plan, incite, or further illegal or unlawful activity or content, such as:\n 1. Violence or terrorism\n 2. Exploitation or harm to children, including the solicitation, creation, acquisition, or dissemination of child exploitative content or failure to report Child Sexual Abuse Material\n 3. Human trafficking, exploitation, and sexual violence\n 4. The illegal distribution of information or materials to minors, including obscene materials, or failure to employ legally required age-gating in connection with such information or materials.\n 5. Sexual solicitation\n 6. Any other criminal activity\n 2. Engage in, promote, incite, or facilitate the harassment, abuse, threatening, or bullying of individuals or groups of individuals\n 3. Engage in, promote, incite, or facilitate discrimination or other unlawful or harmful conduct in the provision of employment, employment benefits, credit, housing, other economic benefits, or other essential goods and services\n 4. Engage in the unauthorized or unlicensed practice of any profession including, but not limited to, financial, legal, medical/health, or related professional practices\n 5. Collect, process, disclose, generate, or infer health, demographic, or other sensitive personal or private information about individuals without rights and consents required by applicable laws\n 6. Engage in or facilitate any action or generate any content that infringes, misappropriates, or otherwise violates any third-party rights, including the outputs or results of any products or services using the Llama Materials\n 7. Create, generate, or facilitate the creation of malicious code, malware, computer viruses or do anything else that could disable, overburden, interfere with or impair the proper working, integrity, operation or appearance of a website or computer system\n2. Engage in, promote, incite, facilitate, or assist in the planning or development of activities that present a risk of death or bodily harm to individuals, including use of Meta Llama 3 related to the following:\n 1. Military, warfare, nuclear industries or applications, espionage, use for materials or activities that are subject to the International Traffic Arms Regulations (ITAR) maintained by the United States Department of State\n 2. Guns and illegal weapons (including weapon development)\n 3. Illegal drugs and regulated/controlled substances\n 4. Operation of critical infrastructure, transportation technologies, or heavy machinery\n 5. Self-harm or harm to others, including suicide, cutting, and eating disorders\n 6. Any content intended to incite or promote violence, abuse, or any infliction of bodily harm to an individual\n3. Intentionally deceive or mislead others, including use of Meta Llama 3 related to the following:\n 1. Generating, promoting, or furthering fraud or the creation or promotion of disinformation\n 2. Generating, promoting, or furthering defamatory content, including the creation of defamatory statements, images, or other content\n 3. Generating, promoting, or further distributing spam\n 4. Impersonating another individual without consent, authorization, or legal right\n 5. Representing that the use of Meta Llama 3 or outputs are human-generated\n 6. Generating or facilitating false online engagement, including fake reviews and other means of fake online engagement\n4. Fail to appropriately disclose to end users any known dangers of your AI system\nPlease report any violation of this Policy, software \u201cbug,\u201d or other problems that could lead to a violation of this Policy through one of the following means:\n * Reporting issues with the model: [https://github.com/meta-llama/llama3](https://github.com/meta-llama/llama3)\n * Reporting risky content generated by the model:\n developers.facebook.com/llama_output_feedback\n * Reporting bugs and security concerns: facebook.com/whitehat/info\n * Reporting violations of the Acceptable Use Policy or unlicensed uses of Meta Llama 3: [email protected]", "extra_gated_fields": {"First Name": "text", "Last Name": "text", "Date of birth": "date_picker", "Country": "country", "Affiliation": "text", "geo": "ip_location", "By clicking Submit below I accept the terms of the license and acknowledge that the information I provide will be collected stored processed and shared in accordance with the Meta Privacy Policy": "checkbox"}, "extra_gated_description": "The information you provide will be collected, stored, processed and shared in accordance with the [Meta Privacy Policy](https://www.facebook.com/privacy/policy/).", "extra_gated_button_content": "Submit", "base_model": "meta-llama/Meta-Llama-3-8B-Instruct", "inference": false, "model_creator": "Meta Llama 3", "model_link": "https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct", "model_type": "llama", "pipeline_tag": "text-generation", "quantized_by": "FriendliAI"}	FriendliAI/Meta-Llama-3-8B-Instruct-fp8	null	[ "transformers", "safetensors", "llama", "text-generation", "facebook", "meta", "pytorch", "llama-3", "conversational", "en", "base_model:meta-llama/Meta-Llama-3-8B-Instruct", "license:other", "autotrain_compatible", "text-generation-inference", "8-bit", "region:us" ]	null	2024-04-19T04:56:43+00:00	[]	[ "en" ]	TAGS #transformers #safetensors #llama #text-generation #facebook #meta #pytorch #llama-3 #conversational #en #base_model-meta-llama/Meta-Llama-3-8B-Instruct #license-other #autotrain_compatible #text-generation-inference #8-bit #region-us	![Friendli Logo](https://i.URL width=) Llama 3 8B Instruct - FP8 ========================= * Model creator: Meta Llama 3 * Original model: Llama 3 8B Instruct Description ----------- This repo contains the Llama 3 8B Instruct model quantized to FP8 by FriendliAI, significantly enhancing its inference efficiency while maintaining high accuracy. Note that FP8 is only supported by NVIDIA Ada, Hopper, and Blackwell GPU architectures. Check out FriendliAI documentation for more details. License ------- Refer to the license of the original model card. Compatibility ------------- This model is compatible with Friendli Container. Prerequisites ------------- * Before you begin, make sure you have signed up for Friendli Suite. You can use Friendli Containers free of charge for four weeks. * Prepare a Personal Access Token following this guide. * Prepare a Friendli Container Secret following this guide. ### Preparing Personal Access Token PAT (Personal Access Token) is the user credential for for logging into our container registry. 1. Sign in Friendli Suite. 2. Go to User Settings > Tokens and click 'Create new token'. 3. Save your created token value. ### Preparing Container Secret Container secret is a credential to launch our Friendli Container images. You should pass the container secret as an environment variable to run the container image. 1. Sign in Friendli Suite. 2. Go to Container > Container Secrets and click 'Create secret'. 3. Save your created secret value. ### Pulling Friendli Container Image 1. Log in to the Docker client using the personal access token created as outlined in this guide. 2. Pull image Running Friendli Container -------------------------- Once you've prepared the image of Friendli Container, you can launch it to create a serving endpoint. --- Original model card: Meta Llama 3 8B Instruct ============================================= Model Details ------------- Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety. Model developers Meta Variations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants. Input Models input text only. Output Models generate text and code only. Model Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. Llama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability. Model Release Date April 18, 2024. Status This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback. License A custom commercial license is available at: URL Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model README. For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go here. Intended Use ------------ Intended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. Out-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English. Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. How to use ---------- This repository contains two versions of Meta-Llama-3-8B-Instruct, for use with transformers and with the original 'llama3' codebase. ### Use with transformers You can run conversational inference using the Transformers pipeline abstraction, or by leveraging the Auto classes with the 'generate()' function. Let's see examples of both. #### Transformers pipeline #### Transformers AutoModelForCausalLM ### Use with 'llama3' Please, follow the instructions in the repository To download Original checkpoints, see the example command below leveraging 'huggingface-cli': For Hugging Face support, we recommend using transformers or TGI, but a similar command works. Hardware and Software --------------------- Training Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute. Carbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program. CO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others. Training Data ------------- Overview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data. Data Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively. Benchmarks ---------- In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here. ### Base pretrained models ### Instruction tuned models ### Responsibility & Safety We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community. Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications. Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience. As part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started. #### Llama 3-Instruct As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case. Safety For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable. Refusals In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2. We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date. #### Responsible release In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision. Misuse If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL #### Critical risks CBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives) We have conducted a two fold assessment of the safety of the model in this area: * Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks. * Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model). ### Cyber Security We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability. ### Child Safety Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences. ### Community Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository. Finally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community. Ethical Considerations and Limitations -------------------------------------- The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress. But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety. Please see the Responsible Use Guide available at URL instructions @article{llama3modelcard, title={Llama 3 Model Card}, author={AI@Meta}, year={2024}, url = {URL } Contributors ------------ Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos	[ "### Preparing Personal Access Token\n\n\nPAT (Personal Access Token) is the user credential for for logging into our container registry.\n\n\n1. Sign in Friendli Suite.\n2. Go to User Settings > Tokens and click 'Create new token'.\n3. Save your created token value.", "### Preparing Container Secret\n\n\nContainer secret is a credential to launch our Friendli Container images.\nYou should pass the container secret as an environment variable to run the container image.\n\n\n1. Sign in Friendli Suite.\n2. Go to Container > Container Secrets and click 'Create secret'.\n3. Save your created secret value.", "### Pulling Friendli Container Image\n\n\n1. Log in to the Docker client using the personal access token created as outlined in this guide.\n2. Pull image\n\n\nRunning Friendli Container\n--------------------------\n\n\nOnce you've prepared the image of Friendli Container, you can launch it to create a serving endpoint.\n\n\n\n\n---\n\n\nOriginal model card: Meta Llama 3 8B Instruct\n=============================================\n\n\nModel Details\n-------------\n\n\nMeta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety.\n\n\nModel developers Meta\n\n\nVariations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants.\n\n\nInput Models input text only.\n\n\nOutput Models generate text and code only.\n\n\nModel Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.\n\n\n\nLlama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability.\n\n\nModel Release Date April 18, 2024.\n\n\nStatus This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback.\n\n\nLicense A custom commercial license is available at: URL\n\n\nWhere to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model README. For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go here.\n\n\nIntended Use\n------------\n\n\nIntended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks.\n\n\nOut-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English.\n\n\nNote: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy.\n\n\nHow to use\n----------\n\n\nThis repository contains two versions of Meta-Llama-3-8B-Instruct, for use with transformers and with the original 'llama3' codebase.", "### Use with transformers\n\n\nYou can run conversational inference using the Transformers pipeline abstraction, or by leveraging the Auto classes with the 'generate()' function. Let's see examples of both.", "#### Transformers pipeline", "#### Transformers AutoModelForCausalLM", "### Use with 'llama3'\n\n\nPlease, follow the instructions in the repository\n\n\nTo download Original checkpoints, see the example command below leveraging 'huggingface-cli':\n\n\nFor Hugging Face support, we recommend using transformers or TGI, but a similar command works.\n\n\nHardware and Software\n---------------------\n\n\nTraining Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.\n\n\nCarbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.\n\n\n\nCO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.\n\n\nTraining Data\n-------------\n\n\nOverview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.\n\n\nData Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.\n\n\nBenchmarks\n----------\n\n\nIn this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here.", "### Base pretrained models", "### Instruction tuned models", "### Responsibility & Safety\n\n\nWe believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.\n\n\nFoundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.\n\n\nRather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.\n\n\nAs part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started.", "#### Llama 3-Instruct\n\n\nAs outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.\n\n\nSafety\n\n\nFor our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.\n\n\nRefusals\n\n\nIn addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.\n\n\nWe built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.", "#### Responsible release\n\n\nIn addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.\n\n\nMisuse\n\n\nIf you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL", "#### Critical risks\n\n\nCBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)\n\n\nWe have conducted a two fold assessment of the safety of the model in this area:\n\n\n* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.\n* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).", "### Cyber Security\n\n\nWe have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability.", "### Child Safety\n\n\nChild Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.", "### Community\n\n\nGenerative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository.\n\n\nFinally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community.\n\n\nEthical Considerations and Limitations\n--------------------------------------\n\n\nThe core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.\n\n\nBut Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.\n\n\nPlease see the Responsible Use Guide available at URL\n\n\ninstructions\n\n\n@article{llama3modelcard,\n\n\ntitle={Llama 3 Model Card},\n\n\nauthor={AI@Meta},\n\n\nyear={2024},\n\n\nurl = {URL\n\n\n}\n\n\nContributors\n------------\n\n\nAaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos" ]	[ "TAGS\n#transformers #safetensors #llama #text-generation #facebook #meta #pytorch #llama-3 #conversational #en #base_model-meta-llama/Meta-Llama-3-8B-Instruct #license-other #autotrain_compatible #text-generation-inference #8-bit #region-us \n", "### Preparing Personal Access Token\n\n\nPAT (Personal Access Token) is the user credential for for logging into our container registry.\n\n\n1. Sign in Friendli Suite.\n2. Go to User Settings > Tokens and click 'Create new token'.\n3. Save your created token value.", "### Preparing Container Secret\n\n\nContainer secret is a credential to launch our Friendli Container images.\nYou should pass the container secret as an environment variable to run the container image.\n\n\n1. Sign in Friendli Suite.\n2. Go to Container > Container Secrets and click 'Create secret'.\n3. Save your created secret value.", "### Pulling Friendli Container Image\n\n\n1. Log in to the Docker client using the personal access token created as outlined in this guide.\n2. Pull image\n\n\nRunning Friendli Container\n--------------------------\n\n\nOnce you've prepared the image of Friendli Container, you can launch it to create a serving endpoint.\n\n\n\n\n---\n\n\nOriginal model card: Meta Llama 3 8B Instruct\n=============================================\n\n\nModel Details\n-------------\n\n\nMeta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety.\n\n\nModel developers Meta\n\n\nVariations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants.\n\n\nInput Models input text only.\n\n\nOutput Models generate text and code only.\n\n\nModel Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.\n\n\n\nLlama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability.\n\n\nModel Release Date April 18, 2024.\n\n\nStatus This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback.\n\n\nLicense A custom commercial license is available at: URL\n\n\nWhere to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model README. For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go here.\n\n\nIntended Use\n------------\n\n\nIntended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks.\n\n\nOut-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English.\n\n\nNote: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy.\n\n\nHow to use\n----------\n\n\nThis repository contains two versions of Meta-Llama-3-8B-Instruct, for use with transformers and with the original 'llama3' codebase.", "### Use with transformers\n\n\nYou can run conversational inference using the Transformers pipeline abstraction, or by leveraging the Auto classes with the 'generate()' function. Let's see examples of both.", "#### Transformers pipeline", "#### Transformers AutoModelForCausalLM", "### Use with 'llama3'\n\n\nPlease, follow the instructions in the repository\n\n\nTo download Original checkpoints, see the example command below leveraging 'huggingface-cli':\n\n\nFor Hugging Face support, we recommend using transformers or TGI, but a similar command works.\n\n\nHardware and Software\n---------------------\n\n\nTraining Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.\n\n\nCarbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.\n\n\n\nCO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.\n\n\nTraining Data\n-------------\n\n\nOverview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.\n\n\nData Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.\n\n\nBenchmarks\n----------\n\n\nIn this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here.", "### Base pretrained models", "### Instruction tuned models", "### Responsibility & Safety\n\n\nWe believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.\n\n\nFoundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.\n\n\nRather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.\n\n\nAs part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started.", "#### Llama 3-Instruct\n\n\nAs outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.\n\n\nSafety\n\n\nFor our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.\n\n\nRefusals\n\n\nIn addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.\n\n\nWe built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.", "#### Responsible release\n\n\nIn addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.\n\n\nMisuse\n\n\nIf you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL", "#### Critical risks\n\n\nCBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)\n\n\nWe have conducted a two fold assessment of the safety of the model in this area:\n\n\n* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.\n* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).", "### Cyber Security\n\n\nWe have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability.", "### Child Safety\n\n\nChild Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.", "### Community\n\n\nGenerative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository.\n\n\nFinally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community.\n\n\nEthical Considerations and Limitations\n--------------------------------------\n\n\nThe core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.\n\n\nBut Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.\n\n\nPlease see the Responsible Use Guide available at URL\n\n\ninstructions\n\n\n@article{llama3modelcard,\n\n\ntitle={Llama 3 Model Card},\n\n\nauthor={AI@Meta},\n\n\nyear={2024},\n\n\nurl = {URL\n\n\n}\n\n\nContributors\n------------\n\n\nAaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos" ]
text-generation	transformers	<!-- header start --> <p align="center"> <img src="https://i.imgur.com/mNM6Cai.png" width="100%" alt="Friendli Logo"> </p> <!-- header end --> # Llama 3 8B - FP8 - Model creator: [Meta Llama 3](https://huggingface.co/meta-llama) - Original model: [Llama 3 8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) ## Description This repo contains the Llama 3 8B model quantized to FP8 by FriendliAI, significantly enhancing its inference efficiency while maintaining high accuracy. Note that FP8 is only supported by NVIDIA Ada, Hopper, and Blackwell GPU architectures. Check out [FriendliAI documentation](https://docs.friendli.ai/) for more details. ## License Refer to the license of the original model card. ## Compatibility This model is compatible with [Friendli Container](https://friendli.ai/products/container/). ## Prerequisites - Before you begin, make sure you have signed up for [Friendli Suite](https://suite.friendli.ai/). You can use Friendli Containers free of charge for four weeks. - Prepare a Personal Access Token following [this guide](#preparing-personal-access-token). - Prepare a Friendli Container Secret following [this guide](#preparing-container-secret). ### Preparing Personal Access Token PAT (Personal Access Token) is the user credential for for logging into our container registry. 1. Sign in [Friendli Suite](https://suite.friendli.ai/). 2. Go to [User Settings > Tokens](https://suite.friendli.ai/user-settings/tokens) and click 'Create new token'. 3. Save your created token value. ### Preparing Container Secret Container secret is a credential to launch our Friendli Container images. You should pass the container secret as an environment variable to run the container image. 1. Sign in [Friendli Suite](https://suite.friendli.ai/). 2. Go to Container > Container Secrets and click 'Create secret'. 3. Save your created secret value. ### Pulling Friendli Container Image 1. Log in to the Docker client using the personal access token created as outlined in [this guide](#preparing-personal-access-token). ```sh export FRIENDLI_PAT="YOUR PAT" docker login registry.friendli.ai -u $YOUR_EMAIL -p $FRIENDLI_PAT ``` 2. Pull image ```sh docker pull registry.friendli.ai/trial ``` ## Running Friendli Container Once you've prepared the image of Friendli Container, you can launch it to create a serving endpoint. ```sh docker run \ --gpus '"device=0"' \ -p 8000:8000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ -e FRIENDLI_CONTAINER_SECRET="YOUR CONTAINER SECRET" \ registry.friendli.ai/trial \ --web-server-port 8000 \ --hf-model-name FriendliAI/Meta-Llama-3-8B-fp8 ``` --- # Original model card: Meta Llama 3 8B ## Model Details Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety. Model developers Meta Variations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants. Input Models input text only. Output Models generate text and code only. Model Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. <table> <tr> <td> </td> <td><strong>Training Data</strong> </td> <td><strong>Params</strong> </td> <td><strong>Context length</strong> </td> <td><strong>GQA</strong> </td> <td><strong>Token count</strong> </td> <td><strong>Knowledge cutoff</strong> </td> </tr> <tr> <td rowspan="2" >Llama 3 </td> <td rowspan="2" >A new mix of publicly available online data. </td> <td>8B </td> <td>8k </td> <td>Yes </td> <td rowspan="2" >15T+ </td> <td>March, 2023 </td> </tr> <tr> <td>70B </td> <td>8k </td> <td>Yes </td> <td>December, 2023 </td> </tr> </table> Llama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability. Model Release Date April 18, 2024. Status This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback. License A custom commercial license is available at: [https://llama.meta.com/llama3/license](https://llama.meta.com/llama3/license) Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model [README](https://github.com/meta-llama/llama3). For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go [here](https://github.com/meta-llama/llama-recipes). ## Intended Use Intended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. Out-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English. Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. ## How to use This repository contains two versions of Meta-Llama-3-8B, for use with transformers and with the original `llama3` codebase. ### Use with transformers See the snippet below for usage with Transformers: ```python >>> import transformers >>> import torch >>> model_id = "meta-llama/Meta-Llama-3-8B" >>> pipeline = transformers.pipeline( "text-generation", model=model_id, model_kwargs={"torch_dtype": torch.bfloat16}, device_map="auto" ) >>> pipeline("Hey how are you doing today?") ``` ### Use with `llama3` Please, follow the instructions in the [repository](https://github.com/meta-llama/llama3). To download Original checkpoints, see the example command below leveraging `huggingface-cli`: ``` huggingface-cli download meta-llama/Meta-Llama-3-8B --include "original/" --local-dir Meta-Llama-3-8B ``` For Hugging Face support, we recommend using transformers or TGI, but a similar command works. ## Hardware and Software Training Factors* We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute. Carbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program. <table> <tr> <td> </td> <td><strong>Time (GPU hours)</strong> </td> <td><strong>Power Consumption (W)</strong> </td> <td><strong>Carbon Emitted(tCO2eq)</strong> </td> </tr> <tr> <td>Llama 3 8B </td> <td>1.3M </td> <td>700 </td> <td>390 </td> </tr> <tr> <td>Llama 3 70B </td> <td>6.4M </td> <td>700 </td> <td>1900 </td> </tr> <tr> <td>Total </td> <td>7.7M </td> <td> </td> <td>2290 </td> </tr> </table> CO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others. ## Training Data Overview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data. Data Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively. ## Benchmarks In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see [here](https://github.com/meta-llama/llama3/blob/main/eval_methodology.md). ### Base pretrained models <table> <tr> <td><strong>Category</strong> </td> <td><strong>Benchmark</strong> </td> <td><strong>Llama 3 8B</strong> </td> <td><strong>Llama2 7B</strong> </td> <td><strong>Llama2 13B</strong> </td> <td><strong>Llama 3 70B</strong> </td> <td><strong>Llama2 70B</strong> </td> </tr> <tr> <td rowspan="6" >General </td> <td>MMLU (5-shot) </td> <td>66.6 </td> <td>45.7 </td> <td>53.8 </td> <td>79.5 </td> <td>69.7 </td> </tr> <tr> <td>AGIEval English (3-5 shot) </td> <td>45.9 </td> <td>28.8 </td> <td>38.7 </td> <td>63.0 </td> <td>54.8 </td> </tr> <tr> <td>CommonSenseQA (7-shot) </td> <td>72.6 </td> <td>57.6 </td> <td>67.6 </td> <td>83.8 </td> <td>78.7 </td> </tr> <tr> <td>Winogrande (5-shot) </td> <td>76.1 </td> <td>73.3 </td> <td>75.4 </td> <td>83.1 </td> <td>81.8 </td> </tr> <tr> <td>BIG-Bench Hard (3-shot, CoT) </td> <td>61.1 </td> <td>38.1 </td> <td>47.0 </td> <td>81.3 </td> <td>65.7 </td> </tr> <tr> <td>ARC-Challenge (25-shot) </td> <td>78.6 </td> <td>53.7 </td> <td>67.6 </td> <td>93.0 </td> <td>85.3 </td> </tr> <tr> <td>Knowledge reasoning </td> <td>TriviaQA-Wiki (5-shot) </td> <td>78.5 </td> <td>72.1 </td> <td>79.6 </td> <td>89.7 </td> <td>87.5 </td> </tr> <tr> <td rowspan="4" >Reading comprehension </td> <td>SQuAD (1-shot) </td> <td>76.4 </td> <td>72.2 </td> <td>72.1 </td> <td>85.6 </td> <td>82.6 </td> </tr> <tr> <td>QuAC (1-shot, F1) </td> <td>44.4 </td> <td>39.6 </td> <td>44.9 </td> <td>51.1 </td> <td>49.4 </td> </tr> <tr> <td>BoolQ (0-shot) </td> <td>75.7 </td> <td>65.5 </td> <td>66.9 </td> <td>79.0 </td> <td>73.1 </td> </tr> <tr> <td>DROP (3-shot, F1) </td> <td>58.4 </td> <td>37.9 </td> <td>49.8 </td> <td>79.7 </td> <td>70.2 </td> </tr> </table> ### Instruction tuned models <table> <tr> <td><strong>Benchmark</strong> </td> <td><strong>Llama 3 8B</strong> </td> <td><strong>Llama 2 7B</strong> </td> <td><strong>Llama 2 13B</strong> </td> <td><strong>Llama 3 70B</strong> </td> <td><strong>Llama 2 70B</strong> </td> </tr> <tr> <td>MMLU (5-shot) </td> <td>68.4 </td> <td>34.1 </td> <td>47.8 </td> <td>82.0 </td> <td>52.9 </td> </tr> <tr> <td>GPQA (0-shot) </td> <td>34.2 </td> <td>21.7 </td> <td>22.3 </td> <td>39.5 </td> <td>21.0 </td> </tr> <tr> <td>HumanEval (0-shot) </td> <td>62.2 </td> <td>7.9 </td> <td>14.0 </td> <td>81.7 </td> <td>25.6 </td> </tr> <tr> <td>GSM-8K (8-shot, CoT) </td> <td>79.6 </td> <td>25.7 </td> <td>77.4 </td> <td>93.0 </td> <td>57.5 </td> </tr> <tr> <td>MATH (4-shot, CoT) </td> <td>30.0 </td> <td>3.8 </td> <td>6.7 </td> <td>50.4 </td> <td>11.6 </td> </tr> </table> ### Responsibility & Safety We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community. Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications. Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience. As part of the Llama 3 release, we updated our [Responsible Use Guide](https://llama.meta.com/responsible-use-guide/) to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including [Meta Llama Guard 2](https://llama.meta.com/purple-llama/) and [Code Shield](https://llama.meta.com/purple-llama/) safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a [reference implementation](https://github.com/meta-llama/llama-recipes/tree/main/recipes/responsible_ai) to get you started. #### Llama 3-Instruct As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case. <span style="text-decoration:underline;">Safety</span> For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable. <span style="text-decoration:underline;">Refusals</span> In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2. We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date. #### Responsible release In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision. Misuse If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy/](https://llama.meta.com/llama3/use-policy/). #### Critical risks <span style="text-decoration:underline;">CBRNE</span> (Chemical, Biological, Radiological, Nuclear, and high yield Explosives) We have conducted a two fold assessment of the safety of the model in this area: * Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks. * Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model). ### <span style="text-decoration:underline;">Cyber Security </span> We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of [equivalent coding capability](https://huggingface.co/spaces/facebook/CyberSecEval). ### <span style="text-decoration:underline;">Child Safety</span> Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences. ### Community Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our [Github repository](https://github.com/meta-llama/PurpleLlama). Finally, we put in place a set of resources including an [output reporting mechanism](https://developers.facebook.com/llama_output_feedback) and [bug bounty program](https://www.facebook.com/whitehat) to continuously improve the Llama technology with the help of the community. ## Ethical Considerations and Limitations The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress. But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating [Purple Llama](https://github.com/facebookresearch/PurpleLlama) solutions into your workflows and specifically [Llama Guard](https://ai.meta.com/research/publications/llama-guard-llm-based-input-output-safeguard-for-human-ai-conversations/) which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety. Please see the Responsible Use Guide available at [http://llama.meta.com/responsible-use-guide](http://llama.meta.com/responsible-use-guide) ## Citation instructions @article{llama3modelcard, title={Llama 3 Model Card}, author={AI@Meta}, year={2024}, url = {https://github.com/meta-llama/llama3/blob/main/MODEL_CARD.md} } ## Contributors Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos	{"language": ["en"], "license": "other", "tags": ["facebook", "meta", "pytorch", "llama", "llama-3"], "model_name": "Llama 3 8B", "license_name": "llama3", "license_link": "LICENSE", "extra_gated_prompt": "### META LLAMA 3 COMMUNITY LICENSE AGREEMENT\nMeta Llama 3 Version Release Date: April 18, 2024\n\"Agreement\" means the terms and conditions for use, reproduction, distribution and modification of the Llama Materials set forth herein.\n\"Documentation\" means the specifications, manuals and documentation accompanying Meta Llama 3 distributed by Meta at https://llama.meta.com/get-started/.\n\"Licensee\" or \"you\" means you, or your employer or any other person or entity (if you are entering into this Agreement on such person or entity\u2019s behalf), of the age required under applicable laws, rules or regulations to provide legal consent and that has legal authority to bind your employer or such other person or entity if you are entering in this Agreement on their behalf.\n\"Meta Llama 3\" means the foundational large language models and software and algorithms, including machine-learning model code, trained model weights, inference-enabling code, training-enabling code, fine-tuning enabling code and other elements of the foregoing distributed by Meta at https://llama.meta.com/llama-downloads.\n\"Llama Materials\" means, collectively, Meta\u2019s proprietary Meta Llama 3 and Documentation (and any portion thereof) made available under this Agreement.\n\"Meta\" or \"we\" means Meta Platforms Ireland Limited (if you are located in or, if you are an entity, your principal place of business is in the EEA or Switzerland) and Meta Platforms, Inc. (if you are located outside of the EEA or Switzerland).\n \n1. License Rights and Redistribution.\na. Grant of Rights. You are granted a non-exclusive, worldwide, non-transferable and royalty-free limited license under Meta\u2019s intellectual property or other rights owned by Meta embodied in the Llama Materials to use, reproduce, distribute, copy, create derivative works of, and make modifications to the Llama Materials.\nb. Redistribution and Use.\ni. If you distribute or make available the Llama Materials (or any derivative works thereof), or a product or service that uses any of them, including another AI model, you shall (A) provide a copy of this Agreement with any such Llama Materials; and (B) prominently display \u201cBuilt with Meta Llama 3\u201d on a related website, user interface, blogpost, about page, or product documentation. If you use the Llama Materials to create, train, fine tune, or otherwise improve an AI model, which is distributed or made available, you shall also include \u201cLlama 3\u201d at the beginning of any such AI model name.\nii. If you receive Llama Materials, or any derivative works thereof, from a Licensee as part of an integrated end user product, then Section 2 of this Agreement will not apply to you.\niii. You must retain in all copies of the Llama Materials that you distribute the following attribution notice within a \u201cNotice\u201d text file distributed as a part of such copies: \u201cMeta Llama 3 is licensed under the Meta Llama 3 Community License, Copyright \u00a9 Meta Platforms, Inc. All Rights Reserved.\u201d\niv. Your use of the Llama Materials must comply with applicable laws and regulations (including trade compliance laws and regulations) and adhere to the Acceptable Use Policy for the Llama Materials (available at https://llama.meta.com/llama3/use-policy), which is hereby incorporated by reference into this Agreement.\nv. You will not use the Llama Materials or any output or results of the Llama Materials to improve any other large language model (excluding Meta Llama 3 or derivative works thereof).\n2. Additional Commercial Terms. If, on the Meta Llama 3 version release date, the monthly active users of the products or services made available by or for Licensee, or Licensee\u2019s affiliates, is greater than 700 million monthly active users in the preceding calendar month, you must request a license from Meta, which Meta may grant to you in its sole discretion, and you are not authorized to exercise any of the rights under this Agreement unless or until Meta otherwise expressly grants you such rights.\n3. Disclaimer of Warranty. UNLESS REQUIRED BY APPLICABLE LAW, THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS THEREFROM ARE PROVIDED ON AN \u201cAS IS\u201d BASIS, WITHOUT WARRANTIES OF ANY KIND, AND META DISCLAIMS ALL WARRANTIES OF ANY KIND, BOTH EXPRESS AND IMPLIED, INCLUDING, WITHOUT LIMITATION, ANY WARRANTIES OF TITLE, NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE. YOU ARE SOLELY RESPONSIBLE FOR DETERMINING THE APPROPRIATENESS OF USING OR REDISTRIBUTING THE LLAMA MATERIALS AND ASSUME ANY RISKS ASSOCIATED WITH YOUR USE OF THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS.\n4. Limitation of Liability. IN NO EVENT WILL META OR ITS AFFILIATES BE LIABLE UNDER ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, TORT, NEGLIGENCE, PRODUCTS LIABILITY, OR OTHERWISE, ARISING OUT OF THIS AGREEMENT, FOR ANY LOST PROFITS OR ANY INDIRECT, SPECIAL, CONSEQUENTIAL, INCIDENTAL, EXEMPLARY OR PUNITIVE DAMAGES, EVEN IF META OR ITS AFFILIATES HAVE BEEN ADVISED OF THE POSSIBILITY OF ANY OF THE FOREGOING.\n5. Intellectual Property.\na. No trademark licenses are granted under this Agreement, and in connection with the Llama Materials, neither Meta nor Licensee may use any name or mark owned by or associated with the other or any of its affiliates, except as required for reasonable and customary use in describing and redistributing the Llama Materials or as set forth in this Section 5(a). Meta hereby grants you a license to use \u201cLlama 3\u201d (the \u201cMark\u201d) solely as required to comply with the last sentence of Section 1.b.i. You will comply with Meta\u2019s brand guidelines (currently accessible at https://about.meta.com/brand/resources/meta/company-brand/ ). All goodwill arising out of your use of the Mark will inure to the benefit of Meta.\nb. Subject to Meta\u2019s ownership of Llama Materials and derivatives made by or for Meta, with respect to any derivative works and modifications of the Llama Materials that are made by you, as between you and Meta, you are and will be the owner of such derivative works and modifications.\nc. If you institute litigation or other proceedings against Meta or any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Llama Materials or Meta Llama 3 outputs or results, or any portion of any of the foregoing, constitutes infringement of intellectual property or other rights owned or licensable by you, then any licenses granted to you under this Agreement shall terminate as of the date such litigation or claim is filed or instituted. You will indemnify and hold harmless Meta from and against any claim by any third party arising out of or related to your use or distribution of the Llama Materials.\n6. Term and Termination. The term of this Agreement will commence upon your acceptance of this Agreement or access to the Llama Materials and will continue in full force and effect until terminated in accordance with the terms and conditions herein. Meta may terminate this Agreement if you are in breach of any term or condition of this Agreement. Upon termination of this Agreement, you shall delete and cease use of the Llama Materials. Sections 3, 4 and 7 shall survive the termination of this Agreement.\n7. Governing Law and Jurisdiction. This Agreement will be governed and construed under the laws of the State of California without regard to choice of law principles, and the UN Convention on Contracts for the International Sale of Goods does not apply to this Agreement. The courts of California shall have exclusive jurisdiction of any dispute arising out of this Agreement.\n### Meta Llama 3 Acceptable Use Policy\nMeta is committed to promoting safe and fair use of its tools and features, including Meta Llama 3. If you access or use Meta Llama 3, you agree to this Acceptable Use Policy (\u201cPolicy\u201d). The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy](https://llama.meta.com/llama3/use-policy)\n#### Prohibited Uses\nWe want everyone to use Meta Llama 3 safely and responsibly. You agree you will not use, or allow others to use, Meta Llama 3 to: 1. Violate the law or others\u2019 rights, including to:\n 1. Engage in, promote, generate, contribute to, encourage, plan, incite, or further illegal or unlawful activity or content, such as:\n 1. Violence or terrorism\n 2. Exploitation or harm to children, including the solicitation, creation, acquisition, or dissemination of child exploitative content or failure to report Child Sexual Abuse Material\n 3. Human trafficking, exploitation, and sexual violence\n 4. The illegal distribution of information or materials to minors, including obscene materials, or failure to employ legally required age-gating in connection with such information or materials.\n 5. Sexual solicitation\n 6. Any other criminal activity\n 2. Engage in, promote, incite, or facilitate the harassment, abuse, threatening, or bullying of individuals or groups of individuals\n 3. Engage in, promote, incite, or facilitate discrimination or other unlawful or harmful conduct in the provision of employment, employment benefits, credit, housing, other economic benefits, or other essential goods and services\n 4. Engage in the unauthorized or unlicensed practice of any profession including, but not limited to, financial, legal, medical/health, or related professional practices\n 5. Collect, process, disclose, generate, or infer health, demographic, or other sensitive personal or private information about individuals without rights and consents required by applicable laws\n 6. Engage in or facilitate any action or generate any content that infringes, misappropriates, or otherwise violates any third-party rights, including the outputs or results of any products or services using the Llama Materials\n 7. Create, generate, or facilitate the creation of malicious code, malware, computer viruses or do anything else that could disable, overburden, interfere with or impair the proper working, integrity, operation or appearance of a website or computer system\n2. Engage in, promote, incite, facilitate, or assist in the planning or development of activities that present a risk of death or bodily harm to individuals, including use of Meta Llama 3 related to the following:\n 1. Military, warfare, nuclear industries or applications, espionage, use for materials or activities that are subject to the International Traffic Arms Regulations (ITAR) maintained by the United States Department of State\n 2. Guns and illegal weapons (including weapon development)\n 3. Illegal drugs and regulated/controlled substances\n 4. Operation of critical infrastructure, transportation technologies, or heavy machinery\n 5. Self-harm or harm to others, including suicide, cutting, and eating disorders\n 6. Any content intended to incite or promote violence, abuse, or any infliction of bodily harm to an individual\n3. Intentionally deceive or mislead others, including use of Meta Llama 3 related to the following:\n 1. Generating, promoting, or furthering fraud or the creation or promotion of disinformation\n 2. Generating, promoting, or furthering defamatory content, including the creation of defamatory statements, images, or other content\n 3. Generating, promoting, or further distributing spam\n 4. Impersonating another individual without consent, authorization, or legal right\n 5. Representing that the use of Meta Llama 3 or outputs are human-generated\n 6. Generating or facilitating false online engagement, including fake reviews and other means of fake online engagement\n4. Fail to appropriately disclose to end users any known dangers of your AI system\nPlease report any violation of this Policy, software \u201cbug,\u201d or other problems that could lead to a violation of this Policy through one of the following means:\n * Reporting issues with the model: [https://github.com/meta-llama/llama3](https://github.com/meta-llama/llama3)\n * Reporting risky content generated by the model:\n developers.facebook.com/llama_output_feedback\n * Reporting bugs and security concerns: facebook.com/whitehat/info\n * Reporting violations of the Acceptable Use Policy or unlicensed uses of Meta Llama 3: [email protected]", "extra_gated_fields": {"First Name": "text", "Last Name": "text", "Date of birth": "date_picker", "Country": "country", "Affiliation": "text", "geo": "ip_location", "By clicking Submit below I accept the terms of the license and acknowledge that the information I provide will be collected stored processed and shared in accordance with the Meta Privacy Policy": "checkbox"}, "extra_gated_description": "The information you provide will be collected, stored, processed and shared in accordance with the [Meta Privacy Policy](https://www.facebook.com/privacy/policy/).", "extra_gated_button_content": "Submit", "base_model": "meta-llama/Meta-Llama-3-8B", "inference": false, "model_creator": "Meta Llama 3", "model_link": "https://huggingface.co/meta-llama/Meta-Llama-3-8B", "model_type": "llama", "pipeline_tag": "text-generation", "quantized_by": "FriendliAI"}	FriendliAI/Meta-Llama-3-8B-fp8	null	[ "transformers", "safetensors", "llama", "text-generation", "facebook", "meta", "pytorch", "llama-3", "conversational", "en", "base_model:meta-llama/Meta-Llama-3-8B", "license:other", "autotrain_compatible", "text-generation-inference", "8-bit", "region:us" ]	null	2024-04-19T04:58:57+00:00	[]	[ "en" ]	TAGS #transformers #safetensors #llama #text-generation #facebook #meta #pytorch #llama-3 #conversational #en #base_model-meta-llama/Meta-Llama-3-8B #license-other #autotrain_compatible #text-generation-inference #8-bit #region-us	![Friendli Logo](https://i.URL width=) Llama 3 8B - FP8 ================ * Model creator: Meta Llama 3 * Original model: Llama 3 8B Description ----------- This repo contains the Llama 3 8B model quantized to FP8 by FriendliAI, significantly enhancing its inference efficiency while maintaining high accuracy. Note that FP8 is only supported by NVIDIA Ada, Hopper, and Blackwell GPU architectures. Check out FriendliAI documentation for more details. License ------- Refer to the license of the original model card. Compatibility ------------- This model is compatible with Friendli Container. Prerequisites ------------- * Before you begin, make sure you have signed up for Friendli Suite. You can use Friendli Containers free of charge for four weeks. * Prepare a Personal Access Token following this guide. * Prepare a Friendli Container Secret following this guide. ### Preparing Personal Access Token PAT (Personal Access Token) is the user credential for for logging into our container registry. 1. Sign in Friendli Suite. 2. Go to User Settings > Tokens and click 'Create new token'. 3. Save your created token value. ### Preparing Container Secret Container secret is a credential to launch our Friendli Container images. You should pass the container secret as an environment variable to run the container image. 1. Sign in Friendli Suite. 2. Go to Container > Container Secrets and click 'Create secret'. 3. Save your created secret value. ### Pulling Friendli Container Image 1. Log in to the Docker client using the personal access token created as outlined in this guide. 2. Pull image Running Friendli Container -------------------------- Once you've prepared the image of Friendli Container, you can launch it to create a serving endpoint. --- Original model card: Meta Llama 3 8B ==================================== Model Details ------------- Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety. Model developers Meta Variations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants. Input Models input text only. Output Models generate text and code only. Model Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. Llama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability. Model Release Date April 18, 2024. Status This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback. License A custom commercial license is available at: URL Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model README. For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go here. Intended Use ------------ Intended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. Out-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English. Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. How to use ---------- This repository contains two versions of Meta-Llama-3-8B, for use with transformers and with the original 'llama3' codebase. ### Use with transformers See the snippet below for usage with Transformers: ### Use with 'llama3' Please, follow the instructions in the repository. To download Original checkpoints, see the example command below leveraging 'huggingface-cli': For Hugging Face support, we recommend using transformers or TGI, but a similar command works. Hardware and Software --------------------- Training Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute. Carbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program. CO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others. Training Data ------------- Overview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data. Data Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively. Benchmarks ---------- In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here. ### Base pretrained models ### Instruction tuned models ### Responsibility & Safety We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community. Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications. Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience. As part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started. #### Llama 3-Instruct As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case. Safety For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable. Refusals In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2. We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date. #### Responsible release In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision. Misuse If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL #### Critical risks CBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives) We have conducted a two fold assessment of the safety of the model in this area: * Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks. * Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model). ### Cyber Security We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability. ### Child Safety Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences. ### Community Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository. Finally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community. Ethical Considerations and Limitations -------------------------------------- The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress. But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety. Please see the Responsible Use Guide available at URL instructions @article{llama3modelcard, title={Llama 3 Model Card}, author={AI@Meta}, year={2024}, url = {URL } Contributors ------------ Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos	[ "### Preparing Personal Access Token\n\n\nPAT (Personal Access Token) is the user credential for for logging into our container registry.\n\n\n1. Sign in Friendli Suite.\n2. Go to User Settings > Tokens and click 'Create new token'.\n3. Save your created token value.", "### Preparing Container Secret\n\n\nContainer secret is a credential to launch our Friendli Container images.\nYou should pass the container secret as an environment variable to run the container image.\n\n\n1. Sign in Friendli Suite.\n2. Go to Container > Container Secrets and click 'Create secret'.\n3. Save your created secret value.", "### Pulling Friendli Container Image\n\n\n1. Log in to the Docker client using the personal access token created as outlined in this guide.\n2. Pull image\n\n\nRunning Friendli Container\n--------------------------\n\n\nOnce you've prepared the image of Friendli Container, you can launch it to create a serving endpoint.\n\n\n\n\n---\n\n\nOriginal model card: Meta Llama 3 8B\n====================================\n\n\nModel Details\n-------------\n\n\nMeta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety.\n\n\nModel developers Meta\n\n\nVariations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants.\n\n\nInput Models input text only.\n\n\nOutput Models generate text and code only.\n\n\nModel Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.\n\n\n\nLlama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability.\n\n\nModel Release Date April 18, 2024.\n\n\nStatus This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback.\n\n\nLicense A custom commercial license is available at: URL\n\n\nWhere to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model README. For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go here.\n\n\nIntended Use\n------------\n\n\nIntended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks.\n\n\nOut-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English.\n\n\nNote: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy.\n\n\nHow to use\n----------\n\n\nThis repository contains two versions of Meta-Llama-3-8B, for use with transformers and with the original 'llama3' codebase.", "### Use with transformers\n\n\nSee the snippet below for usage with Transformers:", "### Use with 'llama3'\n\n\nPlease, follow the instructions in the repository.\n\n\nTo download Original checkpoints, see the example command below leveraging 'huggingface-cli':\n\n\nFor Hugging Face support, we recommend using transformers or TGI, but a similar command works.\n\n\nHardware and Software\n---------------------\n\n\nTraining Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.\n\n\nCarbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.\n\n\n\nCO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.\n\n\nTraining Data\n-------------\n\n\nOverview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.\n\n\nData Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.\n\n\nBenchmarks\n----------\n\n\nIn this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here.", "### Base pretrained models", "### Instruction tuned models", "### Responsibility & Safety\n\n\nWe believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.\n\n\nFoundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.\n\n\nRather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.\n\n\nAs part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started.", "#### Llama 3-Instruct\n\n\nAs outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.\n\n\nSafety\n\n\nFor our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.\n\n\nRefusals\n\n\nIn addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.\n\n\nWe built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.", "#### Responsible release\n\n\nIn addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.\n\n\nMisuse\n\n\nIf you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL", "#### Critical risks\n\n\nCBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)\n\n\nWe have conducted a two fold assessment of the safety of the model in this area:\n\n\n* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.\n* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).", "### Cyber Security\n\n\nWe have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability.", "### Child Safety\n\n\nChild Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.", "### Community\n\n\nGenerative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository.\n\n\nFinally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community.\n\n\nEthical Considerations and Limitations\n--------------------------------------\n\n\nThe core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.\n\n\nBut Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.\n\n\nPlease see the Responsible Use Guide available at URL\n\n\ninstructions\n\n\n@article{llama3modelcard,\n\n\ntitle={Llama 3 Model Card},\n\n\nauthor={AI@Meta},\n\n\nyear={2024},\n\n\nurl = {URL\n\n\n}\n\n\nContributors\n------------\n\n\nAaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos" ]	[ "TAGS\n#transformers #safetensors #llama #text-generation #facebook #meta #pytorch #llama-3 #conversational #en #base_model-meta-llama/Meta-Llama-3-8B #license-other #autotrain_compatible #text-generation-inference #8-bit #region-us \n", "### Preparing Personal Access Token\n\n\nPAT (Personal Access Token) is the user credential for for logging into our container registry.\n\n\n1. Sign in Friendli Suite.\n2. Go to User Settings > Tokens and click 'Create new token'.\n3. Save your created token value.", "### Preparing Container Secret\n\n\nContainer secret is a credential to launch our Friendli Container images.\nYou should pass the container secret as an environment variable to run the container image.\n\n\n1. Sign in Friendli Suite.\n2. Go to Container > Container Secrets and click 'Create secret'.\n3. Save your created secret value.", "### Pulling Friendli Container Image\n\n\n1. Log in to the Docker client using the personal access token created as outlined in this guide.\n2. Pull image\n\n\nRunning Friendli Container\n--------------------------\n\n\nOnce you've prepared the image of Friendli Container, you can launch it to create a serving endpoint.\n\n\n\n\n---\n\n\nOriginal model card: Meta Llama 3 8B\n====================================\n\n\nModel Details\n-------------\n\n\nMeta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety.\n\n\nModel developers Meta\n\n\nVariations Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants.\n\n\nInput Models input text only.\n\n\nOutput Models generate text and code only.\n\n\nModel Architecture Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.\n\n\n\nLlama 3 family of models. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability.\n\n\nModel Release Date April 18, 2024.\n\n\nStatus This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback.\n\n\nLicense A custom commercial license is available at: URL\n\n\nWhere to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model README. For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go here.\n\n\nIntended Use\n------------\n\n\nIntended Use Cases Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks.\n\n\nOut-of-scope Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English.\n\n\nNote: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy.\n\n\nHow to use\n----------\n\n\nThis repository contains two versions of Meta-Llama-3-8B, for use with transformers and with the original 'llama3' codebase.", "### Use with transformers\n\n\nSee the snippet below for usage with Transformers:", "### Use with 'llama3'\n\n\nPlease, follow the instructions in the repository.\n\n\nTo download Original checkpoints, see the example command below leveraging 'huggingface-cli':\n\n\nFor Hugging Face support, we recommend using transformers or TGI, but a similar command works.\n\n\nHardware and Software\n---------------------\n\n\nTraining Factors We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.\n\n\nCarbon Footprint Pretraining utilized a cumulative 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program.\n\n\n\nCO2 emissions during pre-training. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others.\n\n\nTraining Data\n-------------\n\n\nOverview Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.\n\n\nData Freshness The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively.\n\n\nBenchmarks\n----------\n\n\nIn this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see here.", "### Base pretrained models", "### Instruction tuned models", "### Responsibility & Safety\n\n\nWe believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community.\n\n\nFoundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications.\n\n\nRather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience.\n\n\nAs part of the Llama 3 release, we updated our Responsible Use Guide to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including Meta Llama Guard 2 and Code Shield safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a reference implementation to get you started.", "#### Llama 3-Instruct\n\n\nAs outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case.\n\n\nSafety\n\n\nFor our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable.\n\n\nRefusals\n\n\nIn addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2.\n\n\nWe built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date.", "#### Responsible release\n\n\nIn addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision.\n\n\nMisuse\n\n\nIf you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at URL", "#### Critical risks\n\n\nCBRNE (Chemical, Biological, Radiological, Nuclear, and high yield Explosives)\n\n\nWe have conducted a two fold assessment of the safety of the model in this area:\n\n\n* Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks.\n* Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model).", "### Cyber Security\n\n\nWe have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of equivalent coding capability.", "### Child Safety\n\n\nChild Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.", "### Community\n\n\nGenerative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our Github repository.\n\n\nFinally, we put in place a set of resources including an output reporting mechanism and bug bounty program to continuously improve the Llama technology with the help of the community.\n\n\nEthical Considerations and Limitations\n--------------------------------------\n\n\nThe core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.\n\n\nBut Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating Purple Llama solutions into your workflows and specifically Llama Guard which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety.\n\n\nPlease see the Responsible Use Guide available at URL\n\n\ninstructions\n\n\n@article{llama3modelcard,\n\n\ntitle={Llama 3 Model Card},\n\n\nauthor={AI@Meta},\n\n\nyear={2024},\n\n\nurl = {URL\n\n\n}\n\n\nContributors\n------------\n\n\nAaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos" ]
null	null	KVQuant is a methodology for efficient KV cache quantization that incorporates several innovations to acheive accurate low-precision quantization, thereby enabling efficient long context length inference. TLDR: KVQuant addresses the memory bottleneck with long context length inference by quantizing the KV cache to low precision. KVQuant achieves high accuracy with low-precision KV cache quantization by considering several consistent patterns observed in cached KV values across different LLMs, and by developing methods to exploit these patterns, including: - Per-channel, Pre-RoPE Key quantization to better match the outlier channels in Keys - Non-Uniform Quantization (NUQ) to better represent the non-uniform activations - Dense-and-Sparse Quantization to mitigate the impacts of numerical outliers on quantization difficulty - Q-Norm to mitigate distribution shift at ultra low precisions (eg. 2-bit) - Attention-Sink Aware Quantization to avoid quantization error with the first token, which is disproportionately sensitive to quantization error For more details please check out our [paper](https://arxiv.org/abs/2401.18079.pdf). ## Model description Quantizer file for running DBRX with 3-bit KV cache using KVQuant. * Base Model: [DBRX](https://www.databricks.com/blog/introducing-dbrx-new-state-art-open-llm) * Bitwidth: 3-bit * Sparsity Level: 1% ## Links * Paper: [https://arxiv.org/abs/2401.18079.pdf](https://arxiv.org/abs/2401.18079.pdf) * Code: [https://github.com/SqueezeAILab/KVQuant](https://github.com/SqueezeAILab/KVQuant) --- license: mit ---	{}	squeeze-ai-lab/dbrx-base-a3-s1	null	[ "arxiv:2401.18079", "region:us" ]	null	2024-04-19T04:59:02+00:00	[ "2401.18079" ]	[]	TAGS #arxiv-2401.18079 #region-us	KVQuant is a methodology for efficient KV cache quantization that incorporates several innovations to acheive accurate low-precision quantization, thereby enabling efficient long context length inference. TLDR: KVQuant addresses the memory bottleneck with long context length inference by quantizing the KV cache to low precision. KVQuant achieves high accuracy with low-precision KV cache quantization by considering several consistent patterns observed in cached KV values across different LLMs, and by developing methods to exploit these patterns, including: - Per-channel, Pre-RoPE Key quantization to better match the outlier channels in Keys - Non-Uniform Quantization (NUQ) to better represent the non-uniform activations - Dense-and-Sparse Quantization to mitigate the impacts of numerical outliers on quantization difficulty - Q-Norm to mitigate distribution shift at ultra low precisions (eg. 2-bit) - Attention-Sink Aware Quantization to avoid quantization error with the first token, which is disproportionately sensitive to quantization error For more details please check out our paper. ## Model description Quantizer file for running DBRX with 3-bit KV cache using KVQuant. * Base Model: DBRX * Bitwidth: 3-bit * Sparsity Level: 1% ## Links * Paper: URL * Code: URL --- license: mit ---	[ "## Model description\n\nQuantizer file for running DBRX with 3-bit KV cache using KVQuant. \n\n* Base Model: DBRX\n* Bitwidth: 3-bit\n* Sparsity Level: 1%", "## Links\n\n* Paper: URL\n* Code: URL \n\n---\nlicense: mit\n---" ]	[ "TAGS\n#arxiv-2401.18079 #region-us \n", "## Model description\n\nQuantizer file for running DBRX with 3-bit KV cache using KVQuant. \n\n* Base Model: DBRX\n* Bitwidth: 3-bit\n* Sparsity Level: 1%", "## Links\n\n* Paper: URL\n* Code: URL \n\n---\nlicense: mit\n---" ]
null	peft	# Model Card for Model ID <!-- Provide a quick summary of what the model is/does. --> ## Model Details ### Model Description <!-- Provide a longer summary of what this model is. --> - Developed by: [More Information Needed] - Funded by [optional]: [More Information Needed] - Shared by [optional]: [More Information Needed] - Model type: [More Information Needed] - Language(s) (NLP): [More Information Needed] - License: [More Information Needed] - Finetuned from model [optional]: [More Information Needed] ### Model Sources [optional] <!-- Provide the basic links for the model. --> - Repository: [More Information Needed] - Paper [optional]: [More Information Needed] - Demo [optional]: [More Information Needed] ## Uses <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> ### Direct Use <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> [More Information Needed] ### Downstream Use [optional] <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app --> [More Information Needed] ### Out-of-Scope Use <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. --> [More Information Needed] ## Bias, Risks, and Limitations <!-- This section is meant to convey both technical and sociotechnical limitations. --> [More Information Needed] ### Recommendations <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. --> Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. --> [More Information Needed] ### Training Procedure <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. --> #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - Training regime: [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision --> #### Speeds, Sizes, Times [optional] <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. --> [More Information Needed] ## Evaluation <!-- This section describes the evaluation protocols and provides the results. --> ### Testing Data, Factors & Metrics #### Testing Data <!-- This should link to a Dataset Card if possible. --> [More Information Needed] #### Factors <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. --> [More Information Needed] #### Metrics <!-- These are the evaluation metrics being used, ideally with a description of why. --> [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] <!-- Relevant interpretability work for the model goes here --> [More Information Needed] ## Environmental Impact <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly --> Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - Hardware Type: [More Information Needed] - Hours used: [More Information Needed] - Cloud Provider: [More Information Needed] - Compute Region: [More Information Needed] - Carbon Emitted: [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. --> BibTeX: [More Information Needed] APA: [More Information Needed] ## Glossary [optional] <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. --> [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed] ### Framework versions - PEFT 0.10.0	{"library_name": "peft", "base_model": "state-spaces/mamba-1.4b-hf"}	ChlorophyllChampion/Mamba-1.4b-OpenOrca-downsample-0.01	null	[ "peft", "safetensors", "arxiv:1910.09700", "base_model:state-spaces/mamba-1.4b-hf", "region:us" ]	null	2024-04-19T04:59:39+00:00	[ "1910.09700" ]	[]	TAGS #peft #safetensors #arxiv-1910.09700 #base_model-state-spaces/mamba-1.4b-hf #region-us	# Model Card for Model ID ## Model Details ### Model Description - Developed by: - Funded by [optional]: - Shared by [optional]: - Model type: - Language(s) (NLP): - License: - Finetuned from model [optional]: ### Model Sources [optional] - Repository: - Paper [optional]: - Demo [optional]: ## Uses ### Direct Use ### Downstream Use [optional] ### Out-of-Scope Use ## Bias, Risks, and Limitations ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. ## Training Details ### Training Data ### Training Procedure #### Preprocessing [optional] #### Training Hyperparameters - Training regime: #### Speeds, Sizes, Times [optional] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data #### Factors #### Metrics ### Results #### Summary ## Model Examination [optional] ## Environmental Impact Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - Hardware Type: - Hours used: - Cloud Provider: - Compute Region: - Carbon Emitted: ## Technical Specifications [optional] ### Model Architecture and Objective ### Compute Infrastructure #### Hardware #### Software [optional] BibTeX: APA: ## Glossary [optional] ## More Information [optional] ## Model Card Authors [optional] ## Model Card Contact ### Framework versions - PEFT 0.10.0	[ "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\n\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact", "### Framework versions\n\n- PEFT 0.10.0" ]	[ "TAGS\n#peft #safetensors #arxiv-1910.09700 #base_model-state-spaces/mamba-1.4b-hf #region-us \n", "# Model Card for Model ID", "## Model Details", "### Model Description\n\n\n\n\n\n- Developed by: \n- Funded by [optional]: \n- Shared by [optional]: \n- Model type: \n- Language(s) (NLP): \n- License: \n- Finetuned from model [optional]:", "### Model Sources [optional]\n\n\n\n- Repository: \n- Paper [optional]: \n- Demo [optional]:", "## Uses", "### Direct Use", "### Downstream Use [optional]", "### Out-of-Scope Use", "## Bias, Risks, and Limitations", "### Recommendations\n\n\n\nUsers (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.", "## How to Get Started with the Model\n\nUse the code below to get started with the model.", "## Training Details", "### Training Data", "### Training Procedure", "#### Preprocessing [optional]", "#### Training Hyperparameters\n\n- Training regime:", "#### Speeds, Sizes, Times [optional]", "## Evaluation", "### Testing Data, Factors & Metrics", "#### Testing Data", "#### Factors", "#### Metrics", "### Results", "#### Summary", "## Model Examination [optional]", "## Environmental Impact\n\n\n\nCarbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).\n\n- Hardware Type: \n- Hours used: \n- Cloud Provider: \n- Compute Region: \n- Carbon Emitted:", "## Technical Specifications [optional]", "### Model Architecture and Objective", "### Compute Infrastructure", "#### Hardware", "#### Software\n\n\n\n[optional]\n\n\n\nBibTeX:\n\n\n\nAPA:", "## Glossary [optional]", "## More Information [optional]", "## Model Card Authors [optional]", "## Model Card Contact", "### Framework versions\n\n- PEFT 0.10.0" ]

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.