modelId
stringlengths 5
139
| author
stringlengths 2
42
| last_modified
timestamp[us, tz=UTC]date 2020-02-15 11:33:14
2025-08-01 06:28:43
| downloads
int64 0
223M
| likes
int64 0
11.7k
| library_name
stringclasses 546
values | tags
listlengths 1
4.05k
| pipeline_tag
stringclasses 55
values | createdAt
timestamp[us, tz=UTC]date 2022-03-02 23:29:04
2025-08-01 06:27:36
| card
stringlengths 11
1.01M
|
|---|---|---|---|---|---|---|---|---|---|
jphme/Llama-2-13b-chat-german
|
jphme
| 2023-10-06T12:52:01Z | 1,589 | 62 |
transformers
|
[
"transformers",
"pytorch",
"llama",
"text-generation",
"llama-2",
"german",
"deutsch",
"de",
"en",
"dataset:Christoph911/German-legal-SQuAD",
"dataset:philschmid/test_german_squad",
"arxiv:2307.09288",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] |
text-generation
| 2023-07-21T23:01:14Z |
---
language:
- de
- en
pipeline_tag: text-generation
inference: false
tags:
- pytorch
- llama
- llama-2
- german
- deutsch
datasets:
- Christoph911/German-legal-SQuAD
- philschmid/test_german_squad
---
**Please Check out EM German, our new german-speaking LLM model family with significantly improved capabilites. EM German is available in Llama2 7b,13b and 70b and Mistral- and LeoLM-based versions! All information and download links can be found [here](https://github.com/jphme/EM_German/blob/main/README.md).**
# Llama 2 13b Chat German
Llama-2-13b-chat-german is a variant of [Meta](https://huggingface.co/meta-llama)´s [Llama 2 13b Chat](https://huggingface.co/meta-llama/Llama-2-13b-chat-hf) model, finetuned on an additional dataset in German language.
This model is optimized for German text, providing proficiency in understanding, generating, and interacting with German language content. However the model is not yet fully optimized for German language, as it has been trained on a small, experimental dataset and has limited capabilities due to the small parameter count.
Some of the fineunting data is also targeted towards factual retrieval (only answer questions from information in the context and refuse to hallucinate) and the model should perform better for these tasks than original Llama 2 Chat.
I am working on improving the model´s capabilities and will update the model if there is sufficient interest.
A quantized GGML version for use with llama.cpp, kobold.cpp and other GUIs for CPU inference can be found [here](https://huggingface.co/jphme/Llama-2-13b-chat-german-GGML).
Please note the license of the base model, which is contained in the repo under LICENSE.TXT and see the original model card below for more information.
## Data
* Prorietary German Conversation Dataset
* German SQuAD and German legal SQuAD data (see datasets), augmented with "wrong" contexts, to improve factual RAG
## Prompt Template
Llama2 Chat uses a new prompt format:
```
<s>[INST] <<SYS>>
You are a helpful, respectful and honest assistant. Always answer as helpfully as possible, while being safe. Your answers should not include any harmful, unethical, racist, sexist, toxic, dangerous, or illegal content. Please ensure that your responses are socially unbiased and positive in nature.
If a question does not make any sense, or is not factually coherent, explain why instead of answering something not correct. If you don't know the answer to a question, please don't share false information. Please answer in the same language as the user.
<</SYS>>
This is a test question[/INST] This is a answer </s><s>
```
See also the original implementation [here](https://github.com/facebookresearch/llama/blob/main/llama/generation.py#L213).
There is also a (so far non-documented?) method right in transformers to generate the correct tokenization: [LLamaTokenizer._build_conversation_input_ids](https://github.com/huggingface/transformers/blob/b257c46a075419c09e5ce5c5aa39bc346ecdb9a5/src/transformers/models/llama/tokenization_llama.py#L334).
## Results
I did only evaluate the output on a small, handcrafted sample on test prompts in German, confirming that the model's ability to understand and generate German text is above the base model in many situations. More detailed evaluations and examples will be posted soon.
## Examples
*These examples were not in the training set. They have been generated with a quantized version (Q4_0 for the German finetuned version, Q4_1 for the original Llama2 Chat version), so FP16 results should be better.*
### Example 1 (General Knowledge)
**Question**:
> Wie heißt die Hauptstadt von Nordrhein-Westfalen?
**Answer Original Llama 2 13b Chat**:
> Hello! As your helpful and respectful assistant, I'm here to help you with any questions you may have. The main city of North Rhine-Westphalia is called Düsseldorf. Is there anything else you would like to know about this topic or the region in general? Please let me know if there's anything else I can assist you with!
**Answer Llama 2 13b Chat German**:
> Die Hauptstadt von Nordrhein-Westfalen ist Düsseldorf.
### Example 2 (RAG/Factual Retrieval):
**Question**:
> Warum ist die Mongolische Volksrepublik nicht dem Warschauer Pakt beigetreten?
>
> Kontext zur Beantwortung der Frage: Warschauer_Pakt
>
> === Bilaterale Verträge über Freundschaft, Zusammenarbeit und gegenseitigen Beistand === Mit den bilateralen Bündnisverträgen wurde die Verpflichtung zur gegenseitigen Hilfeleistung unterzeichnet, um alle gewaltsamen militärischen Handlungen, die sich gegen die territoriale Integrität und Souveränität einer Vertragspartei richteten, zu verhindern. Den ersten dieser Freundschaftsverträge hatte die Sowjetunion schon während des Krieges am 12. Dezember 1943 mit der tschechoslowakischen Exilregierung abgeschlossen, der am 27. November 1963 für die Tschechoslowakei verlängert wurde. Von 1943 bis 1949 gab es bereits 23 bilaterale Verträge über Freundschaft, Zusammenarbeit und gegenseitigen Beistand (VFZ) der ersten Generation in Osteuropa. Neben diesem Vertragssystem bestanden ab 1956/57 auch weitere Abkommen: * Truppenstationierungsabkommen der Sowjetunion mit der DDR (12. März 1957), * Truppenstationierungsabkommen der Sowjetunion mit der Volksrepublik Polen (17. Dezember 1956), * Truppenstationierungsabkommen der Sowjetunion mit Rumänien (15. April 1957) und * Truppenstationierungsabkommen der Sowjetunion mit Ungarn (27. Mai 1957) jeweils mit einer Laufzeit von 20 Jahren. Aber bereits der Vertrag über die Beziehungen zwischen der DDR und der Sowjetunion vom 20. September 1950 zur Grenzregelung enthielt eine Vereinbarung zur Stationierung von sowjetischen Truppen auf dem Gebiet der DDR. Im Juli 1963 bat auch die Mongolische Volksrepublik, dem Warschauer Pakt nach Artikel 9 des Warschauer Vertrages beizutreten. Dazu hätte ein Sonderprotokoll gefasst werden müssen, da der Vertragstext nach Artikel 4 sich nur auf Europa bezog. Aufgrund des aufkommenden Chinesisch-sowjetischen Zerwürfnisses gab es keinen Beitritt, sondern es blieb bei einem Beobachterstatus. Stattdessen wurden ab 1966 sowjetische Truppenstationierungen vereinbart. Nachdem die Sowjetunion mit der DDR am 12. Juni 1964 einen Vertrag über Freundschaft, Zusammenarbeit und gegenseitigen Beistand schloss, der die volle Einbeziehung der DDR in das bilaterale Bündnissystem vorsah, wurden in der Zeit von 1964 bis 1972 insgesamt 20 Bündnisverträge der zweiten Generation unterzeichnet: Die Hegemonie der Sowjetunion wurde durch die bilateralen Bündnisverträge noch verstärkt, da diese eine unmittelbare Beistandsverpflichtung bei einem bewaffneten Angriff vorsahen, die auch in den meisten Verträgen sich nicht nur auf Europa beschränkte. In der dritten Generation der Verträge über Freundschaft, Zusammenarbeit und gegenseitigen Beistand erfolgte nach der Unterzeichnung der Schlussakte von Helsinki der Konferenz über Sicherheit und Zusammenarbeit in Europa (KSZE) im August 1975 die Einbeziehung der Breschnew-Doktrin und der ökonomischen Integration. Zudem wurde die Laufzeit der Verträge auf 25 Jahre festgelegt. Die DDR unterzeichnete am 7. Oktober 1975 mit der Sowjetunion diesen neuen Vertrag sowie später weitere mit Ungarn (24. März 1977), Polen (29. Mai 1977), Bulgarien (14. September 1977) und der Tschechoslowakei (3. Oktober 1977).
>
> Bitte beantworte die Frage mit einem kurzen, wörtlichen Zitat aus dem Kontext. Sollte der Kontext keine Antwort auf die Frage enthalten, bitte antworte dass die Frage mit den gegebenen Informationen nicht zu beantworten ist.
**Answer Original Llama 2 13b Chat**:
> Die Mongolische Volksrepublik hat nicht dem Warschauer Pakt beigetreten, weil es keinen Beitritt gab, sondern nur einen Beobachterstatus. (Quelle: Der Kontext)
>
**Answer Llama 2 13b Chat German**:
> Aufgrund des aufkommenden Chinesisch-sowjetischen Zerwürfnisses gab es keinen Beitritt, sondern es blieb bei einem Beobachterstatus
### Example 3 (RAG / Factual Retrieval negative):
**Question**:
> Nach was benannte Spielberg seine Produktionsfirma Anfang der 1980er?
>
> Kontext zur Beantwortung der Frage: Webbrowser
>
> == Marktanteile und deren Messung == Bild zeigt die lt. Statistik von StatCounter meistverwendeten Browser nach Ländern 9/2019. Die Statistik für März 2020 ist über folgenden Weblink abrufbar: Die tatsächliche Verbreitung eines Webbrowsers ist nicht zweifelsfrei feststellbar. Verschiedene Anbieter veröffentlichen Statistiken über die Verbreitung von Webbrowsern aufgrund unterschiedlicher häufig recht begrenzter Datenbasen. Da die generelle Verbreitungsrate eines Browsers von verschiedensten Faktoren beeinflusst wird, sind diese Statistiken unterschiedlich aussagekräftig und kommen zu teilweise stark unterschiedlichen, scheinbar widersprüchlichen Ergebnissen. So schwankt die Verbreitung eines Browsers je nach Themengebiet einer aufgerufenen Webseite, Herkunftsregion der aufrufenden Person und dem Zeitpunkt der Messung. Beispielsweise können Benutzer an ihrem Arbeitsplatz an die Verwendung eines vorgegebenen Webbrowsers gebunden sein, privat jedoch einen anderen Browser bevorzugen und verwenden. Auch verschiedene Ereignisse führen zu starken Schwankungen. So steigt der Marktanteil bei der Veröffentlichung einer neuen Hauptversion oder sinkt beim Bekanntwerden einer Sicherheitslücke. Insgesamt geben Browserstatistiken nur die grobe Verteilung und ggf. Trends an, da jede Statistik nur auf der Auswertung der Logdateien ausgewählter Websites beruht und es keine unbestritten repräsentative Auswahl dafür zu verwendender Websites gibt. Die Messungen werden in der Regel anhand des sogenannten ''User-Agent''-Headers durchgeführt, mit der sich der verwendete Browser identifizieren lässt und die dieser bei jeder Anfrage an den Server übermittelt. Bei praktisch allen Browsern kann diese Kennung vom Benutzer verändert werden. Die weltweiten Marktanteile von Desktop-Browsern sehen laut Netmarketshare Stand Juli 2018 und September 2019 prozentual wie folgt aus: Bei Berücksichtigung auch der Browser in den inzwischen sehr häufig verwendeten mobilen Geräten wie Smartphone, Tablet etc. ergibt sich laut den veröffentlichten Statistiken von Netmarketshare, für September 2019 folgendes Bild: Chrome 66 %, Safari 19 %, Firefox 4 %, IE 2,6 %, Edge 2,5 %, QQ 1,4 % etc. Die Verbreitung laut Statistik von StatCounter der verschiedenen Browser auf mobilen Geräten ist hier abrufbar. Die Auswahl der Anzeige ist editierbar.
>
> Bitte beantworte die Frage mit einem kurzen, wörtlichen Zitat aus dem Kontext. Sollte der Kontext keine Antwort auf die Frage enthalten, bitte antworte dass die Frage mit den gegebenen Informationen nicht zu beantworten ist.
**Answer Original Llama 2 13b Chat**:
> Sure! Here's the answer to your question with a direct quote from the context: "According to Netmarketshare, in July 2018 and September 2019, the worldwide market share of desktop browsers looked like this: Chrome 66%, Safari 19%, Firefox 4%, IE 2.6%, Edge 2.5%, QQ 1.4% etc." So, according to Netmarketshare's statistics, in July 2018 and September 2019, Google Chrome had a market share of 66%.
>
**Answer Llama 2 13b Chat German**:
> Mit den gegebenen Informationen ist diese Frage nicht zu beantworten.
----------------------------
----------------------------
# Original **Llama 2** model card
Llama 2 is a collection of pretrained and fine-tuned generative text models ranging in scale from 7 billion to 70 billion parameters. This is the repository for the 13B fine-tuned model, optimized for dialogue use cases and converted for the Hugging Face Transformers format. Links to other models can be found in the index at the bottom.
## Model Details
*Note: Use of this model is governed by the Meta license. In order to download the model weights and tokenizer, please visit the [website](https://ai.meta.com/resources/models-and-libraries/llama-downloads/) and accept our License before requesting access here.*
Meta developed and publicly released the Llama 2 family of large language models (LLMs), a collection of pretrained and fine-tuned generative text models ranging in scale from 7 billion to 70 billion parameters. Our fine-tuned LLMs, called Llama-2-Chat, are optimized for dialogue use cases. Llama-2-Chat models outperform open-source chat models on most benchmarks we tested, and in our human evaluations for helpfulness and safety, are on par with some popular closed-source models like ChatGPT and PaLM.
**Model Developers** Meta
**Variations** Llama 2 comes in a range of parameter sizes — 7B, 13B, and 70B — as well as pretrained and fine-tuned variations.
**Input** Models input text only.
**Output** Models generate text only.
**Model Architecture** Llama 2 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 to human preferences for helpfulness and safety.
||Training Data|Params|Content Length|GQA|Tokens|LR|
|---|---|---|---|---|---|---|
|Llama 2|*A new mix of publicly available online data*|7B|4k|✗|2.0T|3.0 x 10<sup>-4</sup>|
|Llama 2|*A new mix of publicly available online data*|13B|4k|✗|2.0T|3.0 x 10<sup>-4</sup>|
|Llama 2|*A new mix of publicly available online data*|70B|4k|✔|2.0T|1.5 x 10<sup>-4</sup>|
*Llama 2 family of models.* Token counts refer to pretraining data only. All models are trained with a global batch-size of 4M tokens. Bigger models - 70B -- use Grouped-Query Attention (GQA) for improved inference scalability.
**Model Dates** Llama 2 was trained between January 2023 and July 2023.
**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://ai.meta.com/resources/models-and-libraries/llama-downloads/](https://ai.meta.com/resources/models-and-libraries/llama-downloads/)
**Research Paper** ["Llama-2: Open Foundation and Fine-tuned Chat Models"](arxiv.org/abs/2307.09288)
## Intended Use
**Intended Use Cases** Llama 2 is intended for commercial and research use in English. Tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks.
To get the expected features and performance for the chat versions, a specific formatting needs to be followed, including the `INST` and `<<SYS>>` tags, `BOS` and `EOS` tokens, and the whitespaces and breaklines in between (we recommend calling `strip()` on inputs to avoid double-spaces). See our reference code in github for details: [`chat_completion`](https://github.com/facebookresearch/llama/blob/main/llama/generation.py#L212).
**Out-of-scope Uses** Use in any manner that violates applicable laws or regulations (including trade compliance laws).Use in languages other than English. Use in any other way that is prohibited by the Acceptable Use Policy and Licensing Agreement for Llama 2.
## Hardware and Software
**Training Factors** We used custom training libraries, Meta's Research Super Cluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute.
**Carbon Footprint** Pretraining utilized a cumulative 3.3M GPU hours of computation on hardware of type A100-80GB (TDP of 350-400W). Estimated total emissions were 539 tCO2eq, 100% of which were offset by Meta’s sustainability program.
||Time (GPU hours)|Power Consumption (W)|Carbon Emitted(tCO<sub>2</sub>eq)|
|---|---|---|---|
|Llama 2 7B|184320|400|31.22|
|Llama 2 13B|368640|400|62.44|
|Llama 2 70B|1720320|400|291.42|
|Total|3311616||539.00|
**CO<sub>2</sub> emissions during pretraining.** 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 2 was pretrained on 2 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over one million new human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data.
**Data Freshness** The pretraining data has a cutoff of September 2022, but some tuning data is more recent, up to July 2023.
## Evaluation Results
In this section, we report the results for the Llama 1 and Llama 2 models on standard academic benchmarks.For all the evaluations, we use our internal evaluations library.
|Model|Size|Code|Commonsense Reasoning|World Knowledge|Reading Comprehension|Math|MMLU|BBH|AGI Eval|
|---|---|---|---|---|---|---|---|---|---|
|Llama 1|7B|14.1|60.8|46.2|58.5|6.95|35.1|30.3|23.9|
|Llama 1|13B|18.9|66.1|52.6|62.3|10.9|46.9|37.0|33.9|
|Llama 1|33B|26.0|70.0|58.4|67.6|21.4|57.8|39.8|41.7|
|Llama 1|65B|30.7|70.7|60.5|68.6|30.8|63.4|43.5|47.6|
|Llama 2|7B|16.8|63.9|48.9|61.3|14.6|45.3|32.6|29.3|
|Llama 2|13B|24.5|66.9|55.4|65.8|28.7|54.8|39.4|39.1|
|Llama 2|70B|**37.5**|**71.9**|**63.6**|**69.4**|**35.2**|**68.9**|**51.2**|**54.2**|
**Overall performance on grouped academic benchmarks.** *Code:* We report the average pass@1 scores of our models on HumanEval and MBPP. *Commonsense Reasoning:* We report the average of PIQA, SIQA, HellaSwag, WinoGrande, ARC easy and challenge, OpenBookQA, and CommonsenseQA. We report 7-shot results for CommonSenseQA and 0-shot results for all other benchmarks. *World Knowledge:* We evaluate the 5-shot performance on NaturalQuestions and TriviaQA and report the average. *Reading Comprehension:* For reading comprehension, we report the 0-shot average on SQuAD, QuAC, and BoolQ. *MATH:* We report the average of the GSM8K (8 shot) and MATH (4 shot) benchmarks at top 1.
|||TruthfulQA|Toxigen|
|---|---|---|---|
|Llama 1|7B|27.42|23.00|
|Llama 1|13B|41.74|23.08|
|Llama 1|33B|44.19|22.57|
|Llama 1|65B|48.71|21.77|
|Llama 2|7B|33.29|**21.25**|
|Llama 2|13B|41.86|26.10|
|Llama 2|70B|**50.18**|24.60|
**Evaluation of pretrained LLMs on automatic safety benchmarks.** For TruthfulQA, we present the percentage of generations that are both truthful and informative (the higher the better). For ToxiGen, we present the percentage of toxic generations (the smaller the better).
|||TruthfulQA|Toxigen|
|---|---|---|---|
|Llama-2-Chat|7B|57.04|**0.00**|
|Llama-2-Chat|13B|62.18|**0.00**|
|Llama-2-Chat|70B|**64.14**|0.01|
**Evaluation of fine-tuned LLMs on different safety datasets.** Same metric definitions as above.
## Ethical Considerations and Limitations
Llama 2 is a new technology that carries risks with 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 2’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 2, developers should perform safety testing and tuning tailored to their specific applications of the model.
Please see the Responsible Use Guide available at [https://ai.meta.com/llama/responsible-use-guide/](https://ai.meta.com/llama/responsible-use-guide)
## Reporting Issues
Please report any software “bug,” or other problems with the models through one of the following means:
- Reporting issues with the model: [github.com/facebookresearch/llama](http://github.com/facebookresearch/llama)
- Reporting problematic content generated by the model: [developers.facebook.com/llama_output_feedback](http://developers.facebook.com/llama_output_feedback)
- Reporting bugs and security concerns: [facebook.com/whitehat/info](http://facebook.com/whitehat/info)
## Llama Model Index
|Model|Llama2|Llama2-hf|Llama2-chat|Llama2-chat-hf|
|---|---|---|---|---|
|7B| [Link](https://huggingface.co/llamaste/Llama-2-7b) | [Link](https://huggingface.co/llamaste/Llama-2-7b-hf) | [Link](https://huggingface.co/llamaste/Llama-2-7b-chat) | [Link](https://huggingface.co/llamaste/Llama-2-7b-chat-hf)|
|13B| [Link](https://huggingface.co/llamaste/Llama-2-13b) | [Link](https://huggingface.co/llamaste/Llama-2-13b-hf) | [Link](https://huggingface.co/llamaste/Llama-2-13b-chat) | [Link](https://huggingface.co/llamaste/Llama-2-13b-hf)|
|70B| [Link](https://huggingface.co/llamaste/Llama-2-70b) | [Link](https://huggingface.co/llamaste/Llama-2-70b-hf) | [Link](https://huggingface.co/llamaste/Llama-2-70b-chat) | [Link](https://huggingface.co/llamaste/Llama-2-70b-hf)|
|
rickchoi85/mythomax-13b
|
rickchoi85
| 2023-10-06T12:45:21Z | 4 | 0 |
transformers
|
[
"transformers",
"pytorch",
"llama",
"feature-extraction",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
feature-extraction
| 2023-10-06T11:03:39Z |
original model is [Gryphe/MythoMax-L2-13b](https://huggingface.co/Gryphe/MythoMax-L2-13b)
tokenizer fixed for transformers.js
|
rajendrabaskota/tmp_trainer
|
rajendrabaskota
| 2023-10-06T12:41:18Z | 58 | 0 |
transformers
|
[
"transformers",
"pytorch",
"roberta",
"text-classification",
"generated_from_trainer",
"base_model:rajendrabaskota/hc3-wiki-domain-classification-roberta",
"base_model:finetune:rajendrabaskota/hc3-wiki-domain-classification-roberta",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-10-06T12:14:42Z |
---
license: mit
base_model: rajendrabaskota/hc3-wiki-domain-classification-roberta
tags:
- generated_from_trainer
model-index:
- name: tmp_trainer
results: []
---
<!-- 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. -->
# tmp_trainer
This model is a fine-tuned version of [rajendrabaskota/hc3-wiki-domain-classification-roberta](https://huggingface.co/rajendrabaskota/hc3-wiki-domain-classification-roberta) 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: 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: 3.0
### Framework versions
- Transformers 4.33.0
- Pytorch 2.0.0
- Datasets 2.13.0
- Tokenizers 0.13.3
|
ashwincv0112/code-llama-instruction-finetune2
|
ashwincv0112
| 2023-10-06T12:35:20Z | 2 | 0 |
peft
|
[
"peft",
"arxiv:1910.09700",
"base_model:codellama/CodeLlama-7b-Instruct-hf",
"base_model:adapter:codellama/CodeLlama-7b-Instruct-hf",
"region:us"
] | null | 2023-10-06T12:11:54Z |
---
library_name: peft
base_model: codellama/CodeLlama-7b-Instruct-hf
---
# 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]
- **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 Data 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 Data 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]
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- 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 0.6.0.dev0
|
rajendrabaskota/hc3-wiki-domain-classification-roberta
|
rajendrabaskota
| 2023-10-06T12:33:02Z | 112 | 0 |
transformers
|
[
"transformers",
"pytorch",
"roberta",
"text-classification",
"generated_from_trainer",
"base_model:FacebookAI/roberta-base",
"base_model:finetune:FacebookAI/roberta-base",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-10-05T18:04:39Z |
---
license: mit
base_model: roberta-base
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: hc3-wiki-domain-classification-roberta
results: []
---
<!-- 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. -->
# hc3-wiki-domain-classification-roberta
This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1821
- Accuracy: 0.9810
- F1 Score: 0.9810
## 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: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 Score |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|:--------:|
| 0.5388 | 0.04 | 400 | 0.3470 | 0.9626 | 0.9626 |
| 0.4795 | 0.08 | 800 | 0.4603 | 0.9659 | 0.9659 |
| 0.4419 | 0.12 | 1200 | 0.3184 | 0.9622 | 0.9622 |
| 0.3985 | 0.15 | 1600 | 0.3919 | 0.9697 | 0.9697 |
| 0.3954 | 0.19 | 2000 | 0.3571 | 0.9718 | 0.9718 |
| 0.4891 | 0.23 | 2400 | 0.4775 | 0.9668 | 0.9668 |
| 0.4283 | 0.27 | 2800 | 0.3616 | 0.9677 | 0.9677 |
| 0.4157 | 0.31 | 3200 | 0.4152 | 0.9519 | 0.9519 |
| 0.4477 | 0.35 | 3600 | 0.3460 | 0.9673 | 0.9673 |
| 0.426 | 0.39 | 4000 | 0.4334 | 0.9669 | 0.9669 |
| 0.3704 | 0.43 | 4400 | 0.3405 | 0.9634 | 0.9634 |
| 0.4027 | 0.46 | 4800 | 0.3232 | 0.9738 | 0.9738 |
| 0.3704 | 0.5 | 5200 | 0.3475 | 0.9672 | 0.9672 |
| 0.3459 | 0.54 | 5600 | 0.4094 | 0.9738 | 0.9738 |
| 0.3707 | 0.58 | 6000 | 0.3176 | 0.9703 | 0.9703 |
| 0.3145 | 0.62 | 6400 | 0.3329 | 0.9760 | 0.9760 |
| 0.3153 | 0.66 | 6800 | 0.3762 | 0.9733 | 0.9733 |
| 0.293 | 0.7 | 7200 | 0.2815 | 0.9761 | 0.9761 |
| 0.2981 | 0.74 | 7600 | 0.2577 | 0.9771 | 0.9771 |
| 0.2481 | 0.77 | 8000 | 0.2134 | 0.9780 | 0.9780 |
| 0.2418 | 0.81 | 8400 | 0.1978 | 0.9779 | 0.9779 |
| 0.2235 | 0.85 | 8800 | 0.1896 | 0.9794 | 0.9794 |
| 0.1934 | 0.89 | 9200 | 0.1895 | 0.9796 | 0.9796 |
| 0.2167 | 0.93 | 9600 | 0.1804 | 0.9792 | 0.9792 |
| 0.1992 | 0.97 | 10000 | 0.1821 | 0.9810 | 0.9810 |
### Framework versions
- Transformers 4.33.0
- Pytorch 2.0.0
- Datasets 2.13.0
- Tokenizers 0.13.3
|
chrisyuan45/TimeLlama-7b-chat
|
chrisyuan45
| 2023-10-06T12:15:41Z | 54 | 2 |
transformers
|
[
"transformers",
"pytorch",
"llama",
"text-generation",
"arxiv:2310.01074",
"license:cc",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-09-26T12:53:11Z |
---
license: cc
pipeline_tag: text-generation
---
# TimeLlama
TimeLlama is an instruction-finetuned Llama2 series that improves complex temporal reasoning ability.
## Model Details
### Model Description
In this work, we introduce the first multi-source dataset for explainable temporal reasoning, called ExpTime. The dataset contains 26k examples derived from temporal knowledge graph datasets. Each example includes a context with multiple events, a future event to predict, and an explanation for the prediction in the form of temporal reasoning over the events.
To generate the dataset, we propose a novel knowledge-graph-instructed-generation strategy. The dataset supports the comprehensive evaluation of large language models on complex temporal reasoning, future event prediction, and explainability.
Based on ExpTime, we develop TimeLlaMA, a series of LLM models fine-tuned for explainable temporal reasoning. TimeLlaMA builds on the foundation LLM LLaMA-2 and utilizes instruction tuning to follow prompts for making explanations.
### Model Sources
- **Repository:** https://github.com/chenhan97/TimeLlama
- **Paper:** https://arxiv.org/abs/2310.01074
## Uses
### Direct Use
```python
from transformers import LlamaConfig, LlamaTokenizer, LlamaForCausalLM
# Model names: "chrisyuan45/TimeLlama-7b-chat", "chrisyuan45/TimeLlama-13b-chat"
model = LlamaForCausalLM.from_pretrained(
model_name,
return_dict=True,
load_in_8bit=quantization,
device_map="auto",
low_cpu_mem_usage=True)
tokenizer = LlamaTokenizer.from_pretrained(model_name)
```
### Finetune
Please check our repository for the detailed finetuning method.
|
IWR/a2c-PandaReachDense-v3
|
IWR
| 2023-10-06T12:14:34Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"PandaReachDense-v3",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-10-06T12:09:05Z |
---
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: -0.24 +/- 0.15
name: mean_reward
verified: false
---
# **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
...
```
|
softaken/Split-vCard-Tool
|
softaken
| 2023-10-06T12:08:09Z | 0 | 0 | null |
[
"region:us"
] | null | 2023-10-06T12:07:28Z |
Download the Softaken Split vCard Tool for splitting VCF files of any size and creating in any edition. This utility is designed to split a single vCard into multiple cards. This tool has enough capacity to split a single VCF file at a time regardless of file size. Its overall steps are so simple that even non-professional users can operate this application without any impendent. Therefore it holds your imported data quality still after the migration Process. Whereas this program meets with a trouble-free conversion. With the assistance of this software, you can split your desired vCard files. Moreover, you can use its demo version, to evaluate its functions more deeply. Therefore if you find difficulties in working with this tool, then you can go ahead with its technical support team which is active 24/7 to answer your query which is based on the software. In this software, you will get the option to split VCF files into multiple contacts. This will provide an instant view of your VCF contacts data after the splitting. There will be the option to sort the VCF files or contacts according to the attributes. It is an advanced tool which is supporting all editions of the vCard. Split only users' selective vCard contacts for splitting This is an independent application, that performs its functioning without any Outlook installations
Read More:- https://www.softaken.com/split-vcard-pro
|
duwi/rl_course_vizdoom_health_gathering_supreme
|
duwi
| 2023-10-06T12:07:43Z | 0 | 0 |
sample-factory
|
[
"sample-factory",
"tensorboard",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-10-06T12:07:35Z |
---
library_name: sample-factory
tags:
- deep-reinforcement-learning
- reinforcement-learning
- sample-factory
model-index:
- name: APPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: doom_health_gathering_supreme
type: doom_health_gathering_supreme
metrics:
- type: mean_reward
value: 9.65 +/- 3.26
name: mean_reward
verified: false
---
A(n) **APPO** model trained on the **doom_health_gathering_supreme** environment.
This model was trained using Sample-Factory 2.0: https://github.com/alex-petrenko/sample-factory.
Documentation for how to use Sample-Factory can be found at https://www.samplefactory.dev/
## Downloading the model
After installing Sample-Factory, download the model with:
```
python -m sample_factory.huggingface.load_from_hub -r duwi/rl_course_vizdoom_health_gathering_supreme
```
## Using the model
To run the model after download, use the `enjoy` script corresponding to this environment:
```
python -m .usr.local.lib.python3.10.dist-packages.colab_kernel_launcher --algo=APPO --env=doom_health_gathering_supreme --train_dir=./train_dir --experiment=rl_course_vizdoom_health_gathering_supreme
```
You can also upload models to the Hugging Face Hub using the same script with the `--push_to_hub` flag.
See https://www.samplefactory.dev/10-huggingface/huggingface/ for more details
## Training with this model
To continue training with this model, use the `train` script corresponding to this environment:
```
python -m .usr.local.lib.python3.10.dist-packages.colab_kernel_launcher --algo=APPO --env=doom_health_gathering_supreme --train_dir=./train_dir --experiment=rl_course_vizdoom_health_gathering_supreme --restart_behavior=resume --train_for_env_steps=10000000000
```
Note, you may have to adjust `--train_for_env_steps` to a suitably high number as the experiment will resume at the number of steps it concluded at.
|
Ahmeng/audio_classification
|
Ahmeng
| 2023-10-06T12:05:00Z | 159 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"wav2vec2",
"audio-classification",
"generated_from_trainer",
"dataset:minds14",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
audio-classification
| 2023-10-05T11:27:18Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- minds14
metrics:
- accuracy
model-index:
- name: audio_classification
results: []
---
<!-- 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. -->
# audio_classification
This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the minds14 dataset.
It achieves the following results on the evaluation set:
- Loss: 2.6415
- Accuracy: 0.0796
## 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: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| No log | 1.0 | 8 | 2.6461 | 0.0973 |
| No log | 2.0 | 16 | 2.6415 | 0.0796 |
### Framework versions
- Transformers 4.30.2
- Pytorch 1.13.1+cu117
- Datasets 2.13.2
- Tokenizers 0.13.3
|
cointegrated/rubert-tiny-bilingual-nli
|
cointegrated
| 2023-10-06T11:57:57Z | 442 | 8 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"bert",
"text-classification",
"rubert",
"russian",
"nli",
"rte",
"zero-shot-classification",
"ru",
"dataset:cointegrated/nli-rus-translated-v2021",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
zero-shot-classification
| 2022-03-02T23:29:05Z |
---
language: ru
pipeline_tag: zero-shot-classification
tags:
- rubert
- russian
- nli
- rte
- zero-shot-classification
widget:
- text: "Сервис отстойный, кормили невкусно"
candidate_labels: "Мне понравилось, Мне не понравилось"
hypothesis_template: "{}."
datasets:
- cointegrated/nli-rus-translated-v2021
---
# RuBERT-tiny for NLI (natural language inference)
This is the [cointegrated/rubert-tiny](https://huggingface.co/cointegrated/rubert-tiny) model fine-tuned to predict the logical relationship between two short texts: entailment or not entailment.
For more details, see the card for a related model: https://huggingface.co/cointegrated/rubert-base-cased-nli-threeway
|
cointegrated/rubert-base-cased-nli-twoway
|
cointegrated
| 2023-10-06T11:57:41Z | 553 | 6 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"bert",
"text-classification",
"rubert",
"russian",
"nli",
"rte",
"zero-shot-classification",
"ru",
"dataset:cointegrated/nli-rus-translated-v2021",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
zero-shot-classification
| 2022-03-02T23:29:05Z |
---
language: ru
pipeline_tag: zero-shot-classification
tags:
- rubert
- russian
- nli
- rte
- zero-shot-classification
widget:
- text: Я хочу поехать в Австралию
candidate_labels: спорт,путешествия,музыка,кино,книги,наука,политика
hypothesis_template: Тема текста - {}.
datasets:
- cointegrated/nli-rus-translated-v2021
---
# RuBERT for NLI (natural language inference)
This is the [DeepPavlov/rubert-base-cased](https://huggingface.co/DeepPavlov/rubert-base-cased) fine-tuned to predict the logical relationship between two short texts: entailment or not entailment.
For more details, see the card for a similar model: https://huggingface.co/cointegrated/rubert-base-cased-nli-threeway
|
anders0204/ppo-LunarLander-v2
|
anders0204
| 2023-10-06T11:54:34Z | 0 | 0 | null |
[
"tensorboard",
"LunarLander-v2",
"ppo",
"deep-reinforcement-learning",
"reinforcement-learning",
"custom-implementation",
"deep-rl-course",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-09-26T08:58:44Z |
---
tags:
- LunarLander-v2
- ppo
- deep-reinforcement-learning
- reinforcement-learning
- custom-implementation
- deep-rl-course
model-index:
- name: PPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: LunarLander-v2
type: LunarLander-v2
metrics:
- type: mean_reward
value: -127.37 +/- 67.04
name: mean_reward
verified: false
---
# PPO Agent Playing LunarLander-v2
This is a trained model of a PPO agent playing LunarLander-v2.
# Hyperparameters
```python
{'exp_name': 'ppo'
'seed': 1
'torch_deterministic': True
'cuda': True
'track': True
'wandb_project_name': 'cleanRL'
'wandb_entity': None
'capture_video': False
'env_id': 'LunarLander-v2'
'total_timesteps': 5000
'learning_rate': 0.01
'num_envs': 4
'num_steps': 128
'anneal_lr': False
'gae': True
'gamma': 0.999
'gae_lambda': 0.98
'num_minibatches': 8
'update_epochs': 4
'norm_adv': True
'clip_coef': 0.2
'clip_vloss': True
'ent_coef': 0.01
'vf_coef': 0.5
'max_grad_norm': 0.5
'target_kl': None
'repo_id': 'anders0204/ppo-LunarLander-v2'
'batch_size': 512
'minibatch_size': 64}
```
|
Lolimorimorf/damage_trigger_effect_2023-10-06_11_33
|
Lolimorimorf
| 2023-10-06T11:39:18Z | 106 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"token-classification",
"generated_from_trainer",
"base_model:DeepPavlov/rubert-base-cased",
"base_model:finetune:DeepPavlov/rubert-base-cased",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
token-classification
| 2023-10-06T11:38:04Z |
---
base_model: DeepPavlov/rubert-base-cased
tags:
- generated_from_trainer
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: damage_trigger_effect_2023-10-06_11_33
results: []
---
<!-- 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. -->
# damage_trigger_effect_2023-10-06_11_33
This model is a fine-tuned version of [DeepPavlov/rubert-base-cased](https://huggingface.co/DeepPavlov/rubert-base-cased) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3069
- Precision: 0.0
- Recall: 0.0
- F1: 0.0
- Accuracy: 0.9128
## 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: 20
- eval_batch_size: 20
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 15
### Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:---:|:--------:|
| No log | 1.0 | 28 | 0.3496 | 0.0 | 0.0 | 0.0 | 0.9016 |
| No log | 2.0 | 56 | 0.2948 | 0.0 | 0.0 | 0.0 | 0.9147 |
| No log | 3.0 | 84 | 0.2590 | 0.0 | 0.0 | 0.0 | 0.9171 |
| No log | 4.0 | 112 | 0.2689 | 0.0 | 0.0 | 0.0 | 0.9078 |
| No log | 5.0 | 140 | 0.2561 | 0.0 | 0.0 | 0.0 | 0.9101 |
| No log | 6.0 | 168 | 0.2447 | 0.0 | 0.0 | 0.0 | 0.9155 |
| No log | 7.0 | 196 | 0.2621 | 0.0 | 0.0 | 0.0 | 0.9085 |
| No log | 8.0 | 224 | 0.2734 | 0.0 | 0.0 | 0.0 | 0.9143 |
| No log | 9.0 | 252 | 0.2806 | 0.0 | 0.0 | 0.0 | 0.9066 |
| No log | 10.0 | 280 | 0.2954 | 0.0 | 0.0 | 0.0 | 0.9105 |
| No log | 11.0 | 308 | 0.2929 | 0.0 | 0.0 | 0.0 | 0.9128 |
| No log | 12.0 | 336 | 0.2936 | 0.0 | 0.0 | 0.0 | 0.9116 |
| No log | 13.0 | 364 | 0.2948 | 0.0 | 0.0 | 0.0 | 0.9132 |
| No log | 14.0 | 392 | 0.2973 | 0.0 | 0.0 | 0.0 | 0.9151 |
| No log | 15.0 | 420 | 0.3069 | 0.0 | 0.0 | 0.0 | 0.9128 |
### Framework versions
- Transformers 4.34.0
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.14.0
|
Thireus/WizardLM-70B-V1.0-FP32-4.0bpw-h6-exl2
|
Thireus
| 2023-10-06T11:38:53Z | 11 | 0 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"license:llama2",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] |
text-generation
| 2023-09-22T12:08:13Z |
---
inference: false
license: llama2
model_creator: WizardLM
model_link: https://huggingface.co/WizardLM/WizardLM-70B-V1.0
model_name: WizardLM 70B V1.0
model_type: llama
quantized_by: Thireus
---
# WizardLM 70B V1.0 – EXL2
- Model creator: [WizardLM](https://huggingface.co/WizardLM)
- FP32 Original model used for quantization: [WizardLM 70B V1.0](https://huggingface.co/WizardLM/WizardLM-70B-V1.0) – float32
- FP16 Model used for quantization: [WizardLM 70B V1.0-HF](https://huggingface.co/simsim314/WizardLM-70B-V1.0-HF) – float16 of [WizardLM 70B V1.0](https://huggingface.co/WizardLM/WizardLM-70B-V1.0)
- BF16 Model used for quantization: [WizardLM 70B V1.0-BF16](https://huggingface.co/Thireus/WizardLM-70B-V1.0-BF16) – bfloat16 of [WizardLM 70B V1.0](https://huggingface.co/WizardLM/WizardLM-70B-V1.0)
## Models available:
| Link | BITS (-b) | HEAD BITS (-hb) | MEASU-REMENT LENGTH (-ml) | LENGTH (-l) | CAL DATASET (-c) | Size | V. | Max Context Length | Base Model | Layers | VRAM Min*** | VRAM Max*** | PPL** | Comments                                                                                                                         |
| ------ | --------- | --------------- | ------------------------ | ----------- | ---------------- | ---- | ------- | ------------------ | ---- | ---- |------------------ | ------------------ | ------------------ | ---------------------------------------------------------------------------------- |
| [here](https://huggingface.co/Thireus/WizardLM-70B-V1.0-FP32-4.0bpw-h6-exl2/) | 4.0 | 6 | 2048 | 2048 | [0000.parquet](https://huggingface.co/datasets/wikitext/tree/refs%2Fconvert%2Fparquet/wikitext-2-raw-v1/train)* | 33GB | [0.0.2](https://github.com/turboderp/exllamav2/tree/c0dd3412d59c0bc776264512bf76264e954c221d) | 4096 | [FP32](https://huggingface.co/WizardLM/WizardLM-70B-V1.0) | 80 | 39GB | 44GB | 4.15234375 | Good results | | [here](https://huggingface.co/Thireus/WizardLM-70B-V1.0-HF-4.0bpw-h6-exl2/) | 4.0 | 6 | 2048 | 2048 | [0000.parquet](https://huggingface.co/datasets/wikitext/tree/refs%2Fconvert%2Fparquet/wikitext-2-raw-v1/train)* | 35GB | [0.0.1](https://github.com/turboderp/exllamav2/tree/aee7a281708d5faff2ad0ea4b3a3a4b754f458f3) | 4096 | [FP16](https://huggingface.co/simsim314/WizardLM-70B-V1.0-HF) | 80 | 40GB | 44GB | 4.1640625 | Model suffers from poor prompt understanding and logic is affected |
| [here](https://huggingface.co/Thireus/WizardLM-70B-V1.0-BF16-4.0bpw-h6-exl2/) | 4.0 | 6 | 2048 | 2048 | [0000.parquet](https://huggingface.co/datasets/wikitext/tree/refs%2Fconvert%2Fparquet/wikitext-2-raw-v1/train)* | 33GB | [0.0.2](https://github.com/turboderp/exllamav2/tree/ec5164b8a8e282b91aedb2af94dfeb89887656b7) | 4096 | [BF16](https://huggingface.co/Thireus/WizardLM-70B-V1.0-BF16) | 80 | 39GB | 44GB | 4.2421875 | Model suffers from poor prompt understanding and logic is affected |
| [here](https://huggingface.co/Thireus/WizardLM-70B-V1.0-HF-4.0bpw-h8-exl2/) | 4.0 | 8 | 2048 | 2048 | [0000.parquet](https://huggingface.co/datasets/wikitext/tree/refs%2Fconvert%2Fparquet/wikitext-2-raw-v1/train)* | 35GB | [0.0.2](https://github.com/turboderp/exllamav2/tree/a4f2663e310919f007c593030d56ca110f99c261) | 4096 | [FP16](https://huggingface.co/simsim314/WizardLM-70B-V1.0-HF) | 80 | 39GB | 44GB | 4.24609375 | Model suffers from poor prompt understanding and logic is affected |
| [here](https://huggingface.co/Thireus/WizardLM-70B-V1.0-FP32-5.0bpw-h6-exl2/) | 5.0 | 6 | 2048 | 2048 | [0000.parquet](https://huggingface.co/datasets/wikitext/tree/refs%2Fconvert%2Fparquet/wikitext-2-raw-v1/train)* | 41GB | [0.0.2](https://github.com/turboderp/exllamav2/tree/c0dd3412d59c0bc776264512bf76264e954c221d) | 4096 | [FP32](https://huggingface.co/WizardLM/WizardLM-70B-V1.0) | 80 | 47GB | 52GB | 4.06640625 | Best so far. Good results |
| [here](https://huggingface.co/Thireus/WizardLM-70B-V1.0-HF-5.0bpw-h8-exl2/) | 5.0 | 8 | 2048 | 2048 | [0000.parquet](https://huggingface.co/datasets/wikitext/tree/refs%2Fconvert%2Fparquet/wikitext-2-raw-v1/train)* | 44GB | [0.0.2](https://github.com/turboderp/exllamav2/tree/a4f2663e310919f007c593030d56ca110f99c261) | 4096 | [FP16](https://huggingface.co/simsim314/WizardLM-70B-V1.0-HF) | 80 | 48GB | 52GB | 4.09765625 | Model suffers from poor prompt understanding and logic is affected |
| [here](https://huggingface.co/Thireus/WizardLM-70B-V1.0-HF-5.0bpw-h6-exl2/) | 5.0 | 6 | 2048 | 2048 | [0000.parquet](https://huggingface.co/datasets/wikitext/tree/refs%2Fconvert%2Fparquet/wikitext-2-raw-v1/train)* | 44GB | [0.0.1](https://github.com/turboderp/exllamav2/tree/aee7a281708d5faff2ad0ea4b3a3a4b754f458f3) | 4096 | [FP16](https://huggingface.co/simsim314/WizardLM-70B-V1.0-HF) | 80 | 48GB | 52GB | 4.0625 | Model suffers from poor prompt understanding and logic is affected |
| [here](https://huggingface.co/Thireus/WizardLM-70B-V1.0-BF16-5.0bpw-h6-exl2/) | 5.0 | 6 | 2048 | 2048 | [0000.parquet](https://huggingface.co/datasets/wikitext/tree/refs%2Fconvert%2Fparquet/wikitext-2-raw-v1/train)* | 41GB | [0.0.2](https://github.com/turboderp/exllamav2/tree/ec5164b8a8e282b91aedb2af94dfeb89887656b7) | 4096 | [BF16](https://huggingface.co/Thireus/WizardLM-70B-V1.0-BF16) | 80 | 47GB | 52GB | 4.09765625 | Model suffers from poor prompt understanding and logic is affected |
| [here](https://huggingface.co/Thireus/WizardLM-70B-V1.0-HF-6.0bpw-h6-exl2/) | 6.0 | 6 | 2048 | 2048 | [0000.parquet](https://huggingface.co/datasets/wikitext/tree/refs%2Fconvert%2Fparquet/wikitext-2-raw-v1/train)* | 49GB | [0.0.2](https://github.com/turboderp/exllamav2/tree/fae6fb296c6db4e3b1314c49c030541bed98acb9) | 4096 | [FP16](https://huggingface.co/simsim314/WizardLM-70B-V1.0-HF) | 80 | 56GB | 60GB | 4.0703125 | Model suffers from poor prompt understanding and logic is affected |
\* wikitext-2-raw-v1
\*\* Evaluated with text-generation-webui ExLlama v0.0.2 on wikitext-2-raw-v1 (stride 512 and max_length 0). For reference, [TheBloke_WizardLM-70B-V1.0-GPTQ_gptq-4bit-32g-actorder_True](https://huggingface.co/TheBloke/WizardLM-70B-V1.0-GPTQ/tree/gptq-4bit-32g-actorder_True) has a score of 4.1015625 in perplexity.
\*\*\* Without Flash Attention - For better VRAM optimisation, make sure you install https://github.com/Dao-AILab/flash-attention#installation-and-features
## Description:
_This repository contains EXL2 model files for [WizardLM's WizardLM 70B V1.0](https://huggingface.co/WizardLM/WizardLM-70B-V1.0)._
EXL2 is a new format used by ExLlamaV2 – https://github.com/turboderp/exllamav2. EXL2 is based on the same optimization method as GPTQ. The format allows for mixing quantization
levels within a model to achieve any average bitrate between 2 and 8 bits per weight.
## Prompt template (official):
```
A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions. USER: {prompt} ASSISTANT:
```
## Prompt template (suggested):
```
A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions.
USER:
{prompt}
ASSISTANT:
```
## Quantization process:
| Original Model | → | (optional) float16 or bfloat16 Model* | → | Safetensors Model** | → | EXL2 Model |
| -------------- | --- | ------------- | --- | ---------------- | --- | ---------- |
| [WizardLM 70B V1.0](https://huggingface.co/WizardLM/WizardLM-70B-V1.0) | → | [WizardLM 70B V1.0-HF](https://huggingface.co/simsim314/WizardLM-70B-V1.0-HF)* | → | Safetensors** | → | EXL2 |
Example to convert WizardLM-70B-V1.0-HF to EXL2 4.0 bpw with 6-bit head:
```
mkdir -p ~/EXL2/WizardLM-70B-V1.0-HF_4bit # Create the output directory
python convert.py -i ~/float16_safetensored/WizardLM-70B-V1.0-HF -o ~/EXL2/WizardLM-70B-V1.0-HF_4bit -c ~/EXL2/0000.parquet -b 4.0 -hb 6
```
\* Use the following script to convert your local pytorch_model bin files to float16 (you can also choose bfloat16) + safetensors all in one go:
- https://github.com/oobabooga/text-generation-webui/blob/main/convert-to-safetensors.py
(best for sharding and float16/FP16 or bfloat16/BF16 conversion)
Example to convert [WizardLM 70B V1.0](https://huggingface.co/WizardLM/WizardLM-70B-V1.0) directly to float16 safetensors in 10GB shards:
```
python convert-to-safetensors.py ~/original/WizardLM-70B-V1.0 --output ~/float16_safetensored/WizardLM-70B-V1.0 --max-shard-size 10GB
```
Use `--bf16` if you'd like to try bfloat16 instead, but note that there are concerns about quantization quality – https://github.com/turboderp/exllamav2/issues/30#issuecomment-1719009289
\*\* Use any one of the following scripts to convert your local pytorch_model bin files to safetensors:
- https://github.com/turboderp/exllamav2/blob/master/util/convert_safetensors.py (official ExLlamaV2)
- https://huggingface.co/Panchovix/airoboros-l2-70b-gpt4-1.4.1-safetensors/blob/main/bin2safetensors/convert.py (recommended)
- https://gist.github.com/epicfilemcnulty/1f55fd96b08f8d4d6693293e37b4c55e#file-2safetensors-py
## Further reading:
- https://mlabonne.github.io/blog/posts/Introduction_to_Weight_Quantization.html
|
Chaitanya-M/ppo-LunarLander-v2
|
Chaitanya-M
| 2023-10-06T11:35:19Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-10-06T11:34:52Z |
---
library_name: stable-baselines3
tags:
- LunarLander-v2
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: PPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: LunarLander-v2
type: LunarLander-v2
metrics:
- type: mean_reward
value: 224.21 +/- 45.40
name: mean_reward
verified: false
---
# **PPO** Agent playing **LunarLander-v2**
This is a trained model of a **PPO** agent playing **LunarLander-v2**
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
...
```
|
zongxiao/whisper-small-zh-CN
|
zongxiao
| 2023-10-06T11:29:21Z | 94 | 3 |
transformers
|
[
"transformers",
"pytorch",
"whisper",
"automatic-speech-recognition",
"generated_from_trainer",
"zh",
"dataset:mozilla-foundation/common_voice_13_0",
"base_model:openai/whisper-small",
"base_model:finetune:openai/whisper-small",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
automatic-speech-recognition
| 2023-10-05T00:36:01Z |
---
language:
- zh
license: apache-2.0
base_model: openai/whisper-small
tags:
- generated_from_trainer
datasets:
- mozilla-foundation/common_voice_13_0
model-index:
- name: Whisper Small Chinese - zongxiao -1500
results: []
---
<!-- 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. -->
# Whisper Small Chinese - zongxiao -1500
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 13 dataset.
It achieves the following results on the evaluation set:
- eval_loss: 0.3135
- eval_wer_ortho: 73.9671
- eval_wer: 76.1992
- eval_runtime: 14062.1584
- eval_samples_per_second: 0.756
- eval_steps_per_second: 0.378
- epoch: 0.6
- step: 1500
## 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: 2
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 8
- 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: 50
- training_steps: 4000
### Framework versions
- Transformers 4.33.3
- Pytorch 2.0.1
- Datasets 2.14.5
- Tokenizers 0.13.3
|
jake-walker/dqn-SpaceInvadersNoFrameskip-v4
|
jake-walker
| 2023-10-06T11:11:06Z | 1 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"SpaceInvadersNoFrameskip-v4",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-10-06T11:10:28Z |
---
library_name: stable-baselines3
tags:
- SpaceInvadersNoFrameskip-v4
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: DQN
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: SpaceInvadersNoFrameskip-v4
type: SpaceInvadersNoFrameskip-v4
metrics:
- type: mean_reward
value: 669.50 +/- 287.56
name: mean_reward
verified: false
---
# **DQN** Agent playing **SpaceInvadersNoFrameskip-v4**
This is a trained model of a **DQN** agent playing **SpaceInvadersNoFrameskip-v4**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3)
and the [RL Zoo](https://github.com/DLR-RM/rl-baselines3-zoo).
The RL Zoo is a training framework for Stable Baselines3
reinforcement learning agents,
with hyperparameter optimization and pre-trained agents included.
## Usage (with SB3 RL Zoo)
RL Zoo: https://github.com/DLR-RM/rl-baselines3-zoo<br/>
SB3: https://github.com/DLR-RM/stable-baselines3<br/>
SB3 Contrib: https://github.com/Stable-Baselines-Team/stable-baselines3-contrib
Install the RL Zoo (with SB3 and SB3-Contrib):
```bash
pip install rl_zoo3
```
```
# Download model and save it into the logs/ folder
python -m rl_zoo3.load_from_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -orga jake-walker -f logs/
python -m rl_zoo3.enjoy --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
```
If you installed the RL Zoo3 via pip (`pip install rl_zoo3`), from anywhere you can do:
```
python -m rl_zoo3.load_from_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -orga jake-walker -f logs/
python -m rl_zoo3.enjoy --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
```
## Training (with the RL Zoo)
```
python -m rl_zoo3.train --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
# Upload the model and generate video (when possible)
python -m rl_zoo3.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/ -orga jake-walker
```
## Hyperparameters
```python
OrderedDict([('batch_size', 32),
('buffer_size', 100000),
('env_wrapper',
['stable_baselines3.common.atari_wrappers.AtariWrapper']),
('exploration_final_eps', 0.01),
('exploration_fraction', 0.1),
('frame_stack', 4),
('gradient_steps', 1),
('learning_rate', 0.0001),
('learning_starts', 100000),
('n_timesteps', 1000000.0),
('optimize_memory_usage', False),
('policy', 'CnnPolicy'),
('target_update_interval', 1000),
('train_freq', 4),
('normalize', False)])
```
# Environment Arguments
```python
{'render_mode': 'rgb_array'}
```
|
Tensor31/bloom-3b
|
Tensor31
| 2023-10-06T10:58:49Z | 2 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-10-06T10:24:00Z |
---
library_name: peft
---
## Training procedure
### Framework versions
- PEFT 0.5.0
|
duwi/LunarLander-v2
|
duwi
| 2023-10-06T10:41:05Z | 0 | 0 | null |
[
"tensorboard",
"LunarLander-v2",
"ppo",
"deep-reinforcement-learning",
"reinforcement-learning",
"custom-implementation",
"deep-rl-course",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-10-06T10:40:59Z |
---
tags:
- LunarLander-v2
- ppo
- deep-reinforcement-learning
- reinforcement-learning
- custom-implementation
- deep-rl-course
model-index:
- name: PPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: LunarLander-v2
type: LunarLander-v2
metrics:
- type: mean_reward
value: -143.35 +/- 152.05
name: mean_reward
verified: false
---
# PPO Agent Playing LunarLander-v2
This is a trained model of a PPO agent playing LunarLander-v2.
# Hyperparameters
```python
{'exp_name': 'ppo'
'seed': 1
'torch_deterministic': True
'cuda': True
'track': False
'wandb_project_name': 'cleanRL'
'wandb_entity': None
'capture_video': False
'env_id': 'LunarLander-v2'
'total_timesteps': 50000
'learning_rate': 0.00025
'num_envs': 4
'num_steps': 128
'anneal_lr': True
'gae': True
'gamma': 0.99
'gae_lambda': 0.95
'num_minibatches': 4
'update_epochs': 4
'norm_adv': True
'clip_coef': 0.2
'clip_vloss': True
'ent_coef': 0.01
'vf_coef': 0.5
'max_grad_norm': 0.5
'target_kl': None
'repo_id': 'duwi/LunarLander-v2'
'batch_size': 512
'minibatch_size': 128}
```
|
PratikSahu/a2c-PandaReachDense-v3
|
PratikSahu
| 2023-10-06T10:34:03Z | 1 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"PandaReachDense-v3",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-10-06T10:28:27Z |
---
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: -0.18 +/- 0.14
name: mean_reward
verified: false
---
# **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
...
```
|
Evergardener/Learn-t5-large
|
Evergardener
| 2023-10-06T10:26:46Z | 1 | 0 |
peft
|
[
"peft",
"arxiv:1910.09700",
"base_model:google/flan-t5-large",
"base_model:adapter:google/flan-t5-large",
"region:us"
] | null | 2023-09-22T07:21:21Z |
---
library_name: peft
base_model: google/flan-t5-large
---
# 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]
- **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 Data 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 Data 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]
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: True
- load_in_4bit: False
- 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: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
### Framework versions
- PEFT 0.6.0.dev0
|
Trelis/Llama-2-13b-chat-hf-stanford-nil-policy-adapters
|
Trelis
| 2023-10-06T10:13:38Z | 1 | 0 |
peft
|
[
"peft",
"arxiv:1910.09700",
"base_model:meta-llama/Llama-2-13b-chat-hf",
"base_model:adapter:meta-llama/Llama-2-13b-chat-hf",
"region:us"
] | null | 2023-10-06T10:13:31Z |
---
library_name: peft
base_model: meta-llama/Llama-2-13b-chat-hf
---
# 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]
- **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 Data 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 Data 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]
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- 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 0.6.0.dev0
|
ldos/text_shortening_model_v79
|
ldos
| 2023-10-06T10:09:19Z | 103 | 0 |
transformers
|
[
"transformers",
"pytorch",
"t5",
"text2text-generation",
"generated_from_trainer",
"base_model:google-t5/t5-small",
"base_model:finetune:google-t5/t5-small",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-10-06T10:00:59Z |
---
license: apache-2.0
base_model: t5-small
tags:
- generated_from_trainer
model-index:
- name: text_shortening_model_v79
results: []
---
<!-- 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. -->
# text_shortening_model_v79
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.0551
- Bert precision: 0.8947
- Bert recall: 0.8962
- Bert f1-score: 0.895
- Average word count: 6.7804
- Max word count: 16
- Min word count: 1
- Average token count: 10.8466
- % shortened texts with length > 12: 1.5951
## 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: 7e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 40
### Training results
| Training Loss | Epoch | Step | Validation Loss | Bert precision | Bert recall | Bert f1-score | Average word count | Max word count | Min word count | Average token count | % shortened texts with length > 12 |
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:-----------:|:-------------:|:------------------:|:--------------:|:--------------:|:-------------------:|:----------------------------------:|
| 2.0194 | 1.0 | 30 | 1.4487 | 0.8778 | 0.8746 | 0.8755 | 6.7755 | 16 | 1 | 10.7288 | 2.3313 |
| 1.58 | 2.0 | 60 | 1.3193 | 0.8835 | 0.8837 | 0.883 | 6.9301 | 16 | 2 | 10.7791 | 2.3313 |
| 1.4385 | 3.0 | 90 | 1.2492 | 0.8833 | 0.8855 | 0.8839 | 7.0368 | 16 | 2 | 10.9816 | 2.6994 |
| 1.3616 | 4.0 | 120 | 1.2111 | 0.8877 | 0.8873 | 0.887 | 6.8466 | 16 | 2 | 10.7509 | 1.8405 |
| 1.2976 | 5.0 | 150 | 1.1685 | 0.8869 | 0.8878 | 0.8868 | 6.8564 | 17 | 2 | 10.8172 | 1.8405 |
| 1.2495 | 6.0 | 180 | 1.1559 | 0.8885 | 0.8895 | 0.8885 | 6.8577 | 16 | 2 | 10.8564 | 2.0859 |
| 1.201 | 7.0 | 210 | 1.1353 | 0.8889 | 0.891 | 0.8894 | 6.9521 | 16 | 2 | 11.0012 | 2.3313 |
| 1.1717 | 8.0 | 240 | 1.1164 | 0.8892 | 0.89 | 0.8891 | 6.8601 | 16 | 1 | 10.8933 | 2.0859 |
| 1.1352 | 9.0 | 270 | 1.1110 | 0.8902 | 0.8891 | 0.8891 | 6.708 | 16 | 1 | 10.7436 | 1.1043 |
| 1.0984 | 10.0 | 300 | 1.1037 | 0.8901 | 0.8909 | 0.8901 | 6.8233 | 17 | 1 | 10.8503 | 1.9632 |
| 1.0745 | 11.0 | 330 | 1.0937 | 0.8894 | 0.892 | 0.8902 | 6.9362 | 17 | 2 | 10.9742 | 2.3313 |
| 1.0509 | 12.0 | 360 | 1.0907 | 0.8911 | 0.8916 | 0.8908 | 6.8233 | 17 | 1 | 10.8564 | 1.9632 |
| 1.0269 | 13.0 | 390 | 1.0805 | 0.8906 | 0.8934 | 0.8915 | 6.9448 | 17 | 1 | 11.0135 | 2.2086 |
| 1.0126 | 14.0 | 420 | 1.0784 | 0.8912 | 0.8935 | 0.8919 | 6.9264 | 17 | 2 | 10.973 | 2.3313 |
| 0.9959 | 15.0 | 450 | 1.0725 | 0.8929 | 0.8944 | 0.8932 | 6.8294 | 17 | 1 | 10.8957 | 2.2086 |
| 0.9717 | 16.0 | 480 | 1.0715 | 0.8916 | 0.8941 | 0.8924 | 6.919 | 17 | 1 | 10.9963 | 2.0859 |
| 0.9552 | 17.0 | 510 | 1.0727 | 0.8935 | 0.8949 | 0.8937 | 6.8282 | 17 | 1 | 10.9055 | 1.9632 |
| 0.9461 | 18.0 | 540 | 1.0665 | 0.8947 | 0.8955 | 0.8947 | 6.8061 | 17 | 1 | 10.8613 | 1.5951 |
| 0.926 | 19.0 | 570 | 1.0664 | 0.8948 | 0.896 | 0.895 | 6.7853 | 16 | 1 | 10.8515 | 1.3497 |
| 0.9192 | 20.0 | 600 | 1.0636 | 0.8948 | 0.8953 | 0.8946 | 6.7718 | 16 | 1 | 10.8209 | 1.4724 |
| 0.9101 | 21.0 | 630 | 1.0581 | 0.8954 | 0.897 | 0.8957 | 6.8221 | 16 | 1 | 10.8724 | 1.5951 |
| 0.899 | 22.0 | 660 | 1.0599 | 0.8954 | 0.8974 | 0.8959 | 6.8405 | 16 | 1 | 10.8982 | 1.5951 |
| 0.8843 | 23.0 | 690 | 1.0586 | 0.8943 | 0.8962 | 0.8948 | 6.8393 | 17 | 2 | 10.9055 | 1.9632 |
| 0.8779 | 24.0 | 720 | 1.0572 | 0.8932 | 0.8961 | 0.8942 | 6.8736 | 17 | 2 | 10.9656 | 2.0859 |
| 0.8725 | 25.0 | 750 | 1.0573 | 0.8939 | 0.8963 | 0.8947 | 6.8098 | 16 | 2 | 10.9104 | 1.7178 |
| 0.8567 | 26.0 | 780 | 1.0591 | 0.8951 | 0.8968 | 0.8955 | 6.7926 | 17 | 1 | 10.8945 | 1.5951 |
| 0.8549 | 27.0 | 810 | 1.0577 | 0.8945 | 0.8962 | 0.8948 | 6.8135 | 17 | 1 | 10.9018 | 1.8405 |
| 0.8467 | 28.0 | 840 | 1.0570 | 0.8948 | 0.8961 | 0.895 | 6.7669 | 16 | 1 | 10.8405 | 1.4724 |
| 0.833 | 29.0 | 870 | 1.0577 | 0.895 | 0.896 | 0.895 | 6.7546 | 16 | 1 | 10.8294 | 1.3497 |
| 0.8284 | 30.0 | 900 | 1.0548 | 0.8942 | 0.8957 | 0.8945 | 6.7816 | 16 | 1 | 10.8589 | 1.4724 |
| 0.8296 | 31.0 | 930 | 1.0565 | 0.8947 | 0.8967 | 0.8952 | 6.8037 | 16 | 1 | 10.8982 | 1.4724 |
| 0.8156 | 32.0 | 960 | 1.0550 | 0.8945 | 0.8961 | 0.8948 | 6.7914 | 16 | 2 | 10.8601 | 1.5951 |
| 0.8095 | 33.0 | 990 | 1.0567 | 0.8944 | 0.8962 | 0.8948 | 6.8049 | 16 | 2 | 10.881 | 1.7178 |
| 0.8066 | 34.0 | 1020 | 1.0564 | 0.8948 | 0.8961 | 0.895 | 6.7853 | 16 | 1 | 10.8405 | 1.8405 |
| 0.817 | 35.0 | 1050 | 1.0567 | 0.8951 | 0.8961 | 0.8952 | 6.7509 | 16 | 1 | 10.8172 | 1.5951 |
| 0.8155 | 36.0 | 1080 | 1.0563 | 0.8949 | 0.8964 | 0.8952 | 6.7669 | 16 | 1 | 10.838 | 1.5951 |
| 0.808 | 37.0 | 1110 | 1.0560 | 0.8946 | 0.8965 | 0.8951 | 6.7926 | 16 | 1 | 10.8675 | 1.7178 |
| 0.8049 | 38.0 | 1140 | 1.0554 | 0.895 | 0.8965 | 0.8953 | 6.7742 | 16 | 1 | 10.8393 | 1.4724 |
| 0.8002 | 39.0 | 1170 | 1.0550 | 0.8946 | 0.8962 | 0.8949 | 6.7877 | 16 | 1 | 10.8491 | 1.5951 |
| 0.7912 | 40.0 | 1200 | 1.0551 | 0.8947 | 0.8962 | 0.895 | 6.7804 | 16 | 1 | 10.8466 | 1.5951 |
### Framework versions
- Transformers 4.33.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.13.3
|
takumi12/id2pg_pattern1_ja_b4e20_test
|
takumi12
| 2023-10-06T10:01:49Z | 2 | 0 |
peft
|
[
"peft",
"arxiv:1910.09700",
"base_model:codellama/CodeLlama-13b-Instruct-hf",
"base_model:adapter:codellama/CodeLlama-13b-Instruct-hf",
"region:us"
] | null | 2023-10-06T10:01:41Z |
---
library_name: peft
base_model: codellama/CodeLlama-13b-Instruct-hf
---
# 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]
- **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 Data 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 Data 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]
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: True
- load_in_4bit: False
- 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: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
### Framework versions
- PEFT 0.6.0.dev0
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: True
- load_in_4bit: False
- 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: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
### Framework versions
- PEFT 0.6.0.dev0
|
ShaunThayil/training-1
|
ShaunThayil
| 2023-10-06T10:01:28Z | 181 | 0 |
transformers
|
[
"transformers",
"pytorch",
"roberta",
"text-classification",
"generated_from_trainer",
"base_model:FacebookAI/roberta-base",
"base_model:finetune:FacebookAI/roberta-base",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-09-18T12:47:35Z |
---
license: mit
base_model: roberta-base
tags:
- generated_from_trainer
metrics:
- accuracy
- precision
- recall
- f1
model-index:
- name: training-1
results: []
---
<!-- 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. -->
# training-1
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: 0.0448
- Accuracy: 0.9937
- Precision: 0.9912
- Recall: 0.9859
- F1: 0.9885
## 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: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|
| No log | 0.5 | 302 | 0.0546 | 0.9870 | 0.9737 | 0.9789 | 0.9763 |
| No log | 1.0 | 604 | 0.0511 | 0.9913 | 0.9911 | 0.9771 | 0.9840 |
| 0.1032 | 1.5 | 906 | 0.0558 | 0.9899 | 0.9807 | 0.9824 | 0.9815 |
| 0.1032 | 2.0 | 1208 | 0.0467 | 0.9928 | 0.9982 | 0.9754 | 0.9866 |
| 0.0353 | 2.5 | 1510 | 0.0411 | 0.9937 | 0.9929 | 0.9842 | 0.9885 |
| 0.0353 | 3.0 | 1812 | 0.0460 | 0.9932 | 0.9911 | 0.9842 | 0.9876 |
| 0.0183 | 3.49 | 2114 | 0.0423 | 0.9937 | 0.9947 | 0.9824 | 0.9885 |
| 0.0183 | 3.99 | 2416 | 0.0476 | 0.9932 | 0.9911 | 0.9842 | 0.9876 |
| 0.013 | 4.49 | 2718 | 0.0463 | 0.9932 | 0.9911 | 0.9842 | 0.9876 |
| 0.013 | 4.99 | 3020 | 0.0448 | 0.9937 | 0.9912 | 0.9859 | 0.9885 |
### Framework versions
- Transformers 4.33.1
- Pytorch 2.2.0.dev20230913+cu121
- Datasets 2.14.5
- Tokenizers 0.13.3
|
jiang9527li/SpaceInvadersNoFrameskip-v4
|
jiang9527li
| 2023-10-06T09:51:10Z | 1 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"SpaceInvadersNoFrameskip-v4",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-10-06T09:50:37Z |
---
library_name: stable-baselines3
tags:
- SpaceInvadersNoFrameskip-v4
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: DQN
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: SpaceInvadersNoFrameskip-v4
type: SpaceInvadersNoFrameskip-v4
metrics:
- type: mean_reward
value: 525.50 +/- 99.31
name: mean_reward
verified: false
---
# **DQN** Agent playing **SpaceInvadersNoFrameskip-v4**
This is a trained model of a **DQN** agent playing **SpaceInvadersNoFrameskip-v4**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3)
and the [RL Zoo](https://github.com/DLR-RM/rl-baselines3-zoo).
The RL Zoo is a training framework for Stable Baselines3
reinforcement learning agents,
with hyperparameter optimization and pre-trained agents included.
## Usage (with SB3 RL Zoo)
RL Zoo: https://github.com/DLR-RM/rl-baselines3-zoo<br/>
SB3: https://github.com/DLR-RM/stable-baselines3<br/>
SB3 Contrib: https://github.com/Stable-Baselines-Team/stable-baselines3-contrib
Install the RL Zoo (with SB3 and SB3-Contrib):
```bash
pip install rl_zoo3
```
```
# Download model and save it into the logs/ folder
python -m rl_zoo3.load_from_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -orga jiang9527li -f logs/
python -m rl_zoo3.enjoy --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
```
If you installed the RL Zoo3 via pip (`pip install rl_zoo3`), from anywhere you can do:
```
python -m rl_zoo3.load_from_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -orga jiang9527li -f logs/
python -m rl_zoo3.enjoy --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
```
## Training (with the RL Zoo)
```
python -m rl_zoo3.train --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
# Upload the model and generate video (when possible)
python -m rl_zoo3.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/ -orga jiang9527li
```
## Hyperparameters
```python
OrderedDict([('batch_size', 32),
('buffer_size', 100000),
('env_wrapper',
['stable_baselines3.common.atari_wrappers.AtariWrapper']),
('exploration_final_eps', 0.01),
('exploration_fraction', 0.1),
('frame_stack', 4),
('gradient_steps', 1),
('learning_rate', 0.0001),
('learning_starts', 100000),
('n_timesteps', 1000000.0),
('optimize_memory_usage', False),
('policy', 'CnnPolicy'),
('target_update_interval', 1000),
('train_freq', 4),
('normalize', False)])
```
# Environment Arguments
```python
{'render_mode': 'rgb_array'}
```
|
etri-xainlp/llama2-ko-13b-instruct
|
etri-xainlp
| 2023-10-06T09:50:41Z | 2,237 | 1 |
transformers
|
[
"transformers",
"pytorch",
"llama",
"text-generation",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-10-06T04:43:16Z |
---
license: apache-2.0
---
# llama2-ko-13b-instruct
This model is a fine-tuned version of [meta-llama/Llama-13b-hf](https://huggingface.co/meta-llama/Llama-2-13b-hf) on an instruction-following dataset(650k).
|
EladAssia/rl_course_vizdoom_health_gathering_supreme
|
EladAssia
| 2023-10-06T09:49:05Z | 0 | 0 |
sample-factory
|
[
"sample-factory",
"tensorboard",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-10-06T09:48:59Z |
---
library_name: sample-factory
tags:
- deep-reinforcement-learning
- reinforcement-learning
- sample-factory
model-index:
- name: APPO
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: doom_health_gathering_supreme
type: doom_health_gathering_supreme
metrics:
- type: mean_reward
value: 10.43 +/- 4.32
name: mean_reward
verified: false
---
A(n) **APPO** model trained on the **doom_health_gathering_supreme** environment.
This model was trained using Sample-Factory 2.0: https://github.com/alex-petrenko/sample-factory.
Documentation for how to use Sample-Factory can be found at https://www.samplefactory.dev/
## Downloading the model
After installing Sample-Factory, download the model with:
```
python -m sample_factory.huggingface.load_from_hub -r EladAssia/rl_course_vizdoom_health_gathering_supreme
```
## Using the model
To run the model after download, use the `enjoy` script corresponding to this environment:
```
python -m .usr.local.lib.python3.10.dist-packages.colab_kernel_launcher --algo=APPO --env=doom_health_gathering_supreme --train_dir=./train_dir --experiment=rl_course_vizdoom_health_gathering_supreme
```
You can also upload models to the Hugging Face Hub using the same script with the `--push_to_hub` flag.
See https://www.samplefactory.dev/10-huggingface/huggingface/ for more details
## Training with this model
To continue training with this model, use the `train` script corresponding to this environment:
```
python -m .usr.local.lib.python3.10.dist-packages.colab_kernel_launcher --algo=APPO --env=doom_health_gathering_supreme --train_dir=./train_dir --experiment=rl_course_vizdoom_health_gathering_supreme --restart_behavior=resume --train_for_env_steps=10000000000
```
Note, you may have to adjust `--train_for_env_steps` to a suitably high number as the experiment will resume at the number of steps it concluded at.
|
gianma/ludwig-webinar
|
gianma
| 2023-10-06T09:46:32Z | 0 | 0 |
peft
|
[
"peft",
"base_model:meta-llama/Llama-2-7b-hf",
"base_model:adapter:meta-llama/Llama-2-7b-hf",
"region:us"
] | null | 2023-09-03T17:46:47Z |
---
library_name: peft
base_model: meta-llama/Llama-2-7b-hf
---
## 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: float16
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: float16
### Framework versions
- PEFT 0.5.0
- PEFT 0.5.0
|
NannyML/amazon-reviews-sentiment-bert-base-uncased-6000-samples
|
NannyML
| 2023-10-06T09:36:25Z | 110 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"text-classification",
"generated_from_trainer",
"dataset:amazon_reviews_multi",
"base_model:nlptown/bert-base-multilingual-uncased-sentiment",
"base_model:finetune:nlptown/bert-base-multilingual-uncased-sentiment",
"license:mit",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-09-28T14:18:01Z |
---
license: mit
base_model: nlptown/bert-base-multilingual-uncased-sentiment
tags:
- generated_from_trainer
datasets:
- amazon_reviews_multi
metrics:
- accuracy
- f1
model-index:
- name: amazon-reviews-sentiment-bert-base-uncased-6000-samples
results:
- task:
name: Text Classification
type: text-classification
dataset:
name: amazon_reviews_multi
type: amazon_reviews_multi
config: en
split: validation
args: en
metrics:
- name: Accuracy
type: accuracy
value: 0.7678571428571429
- name: F1
type: f1
value: 0.7167992873886065
---
<!-- 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. -->
# amazon-reviews-sentiment-bert-base-uncased-6000-samples
This model is a fine-tuned version of [nlptown/bert-base-multilingual-uncased-sentiment](https://huggingface.co/nlptown/bert-base-multilingual-uncased-sentiment) on the amazon_reviews_multi dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5890
- Accuracy: 0.7679
- F1: 0.7168
## Predicted labels
- LABEL_0: Negative review
- LABEL_1: Neutral review
- LABEL_2: Positive review
## 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: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| No log | 1.0 | 188 | 0.5745 | 0.7586 | 0.7149 |
| No log | 2.0 | 376 | 0.5890 | 0.7679 | 0.7168 |
### Framework versions
- Transformers 4.33.2
- Pytorch 2.0.0
- Datasets 2.14.6.dev0
- Tokenizers 0.13.3
|
idrishaidar/Multilingual-MiniLM-L12-H384-en-id
|
idrishaidar
| 2023-10-06T09:35:17Z | 11 | 0 |
sentence-transformers
|
[
"sentence-transformers",
"pytorch",
"bert",
"feature-extraction",
"sentence-similarity",
"transformers",
"autotrain_compatible",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] |
sentence-similarity
| 2023-10-06T09:33:27Z |
---
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
---
# {MODEL_NAME}
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search.
<!--- Describe your model here -->
## Usage (Sentence-Transformers)
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
```
pip install -U sentence-transformers
```
Then you can use the model like this:
```python
from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]
model = SentenceTransformer('{MODEL_NAME}')
embeddings = model.encode(sentences)
print(embeddings)
```
## Usage (HuggingFace Transformers)
Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.
```python
from transformers import AutoTokenizer, AutoModel
import torch
#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
token_embeddings = model_output[0] #First element of model_output contains all token embeddings
input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']
# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('{MODEL_NAME}')
model = AutoModel.from_pretrained('{MODEL_NAME}')
# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
# Compute token embeddings
with torch.no_grad():
model_output = model(**encoded_input)
# Perform pooling. In this case, mean pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
print("Sentence embeddings:")
print(sentence_embeddings)
```
## Evaluation Results
<!--- Describe how your model was evaluated -->
For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name={MODEL_NAME})
## Training
The model was trained with the parameters:
**DataLoader**:
`torch.utils.data.dataloader.DataLoader` of length 969 with parameters:
```
{'batch_size': 32, 'sampler': 'torch.utils.data.sampler.RandomSampler', 'batch_sampler': 'torch.utils.data.sampler.BatchSampler'}
```
**Loss**:
`sentence_transformers.losses.MSELoss.MSELoss`
Parameters of the fit()-Method:
```
{
"epochs": 10,
"evaluation_steps": 0,
"evaluator": "NoneType",
"max_grad_norm": 1,
"optimizer_class": "<class 'torch.optim.adamw.AdamW'>",
"optimizer_params": {
"lr": 2e-05
},
"scheduler": "WarmupLinear",
"steps_per_epoch": null,
"warmup_steps": 969,
"weight_decay": 0.01
}
```
## Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel
(1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
)
```
## Citing & Authors
<!--- Describe where people can find more information -->
|
ldos/text_shortening_model_v77
|
ldos
| 2023-10-06T09:23:59Z | 104 | 0 |
transformers
|
[
"transformers",
"pytorch",
"t5",
"text2text-generation",
"generated_from_trainer",
"base_model:google-t5/t5-small",
"base_model:finetune:google-t5/t5-small",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-10-06T09:15:27Z |
---
license: apache-2.0
base_model: t5-small
tags:
- generated_from_trainer
model-index:
- name: text_shortening_model_v77
results: []
---
<!-- 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. -->
# text_shortening_model_v77
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.4183
- Bert precision: 0.8989
- Bert recall: 0.9008
- Bert f1-score: 0.8994
- Average word count: 6.9571
- Max word count: 15
- Min word count: 2
- Average token count: 11.2896
- % shortened texts with length > 12: 2.0859
## 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: 0.0003
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 40
### Training results
| Training Loss | Epoch | Step | Validation Loss | Bert precision | Bert recall | Bert f1-score | Average word count | Max word count | Min word count | Average token count | % shortened texts with length > 12 |
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:-----------:|:-------------:|:------------------:|:--------------:|:--------------:|:-------------------:|:----------------------------------:|
| 1.6557 | 1.0 | 30 | 1.2281 | 0.8905 | 0.8844 | 0.887 | 6.5926 | 15 | 1 | 10.4699 | 1.4724 |
| 1.2524 | 2.0 | 60 | 1.1200 | 0.8933 | 0.8902 | 0.8913 | 6.7227 | 15 | 2 | 10.8294 | 0.9816 |
| 1.1024 | 3.0 | 90 | 1.0914 | 0.8952 | 0.8931 | 0.8937 | 6.7706 | 16 | 2 | 10.811 | 1.1043 |
| 0.9921 | 4.0 | 120 | 1.0805 | 0.8935 | 0.8943 | 0.8935 | 6.9141 | 17 | 2 | 11.0454 | 1.3497 |
| 0.8865 | 5.0 | 150 | 1.1025 | 0.8971 | 0.8949 | 0.8956 | 6.7607 | 16 | 2 | 10.9828 | 1.4724 |
| 0.8273 | 6.0 | 180 | 1.1039 | 0.9005 | 0.8989 | 0.8993 | 6.7583 | 16 | 2 | 10.8982 | 1.9632 |
| 0.7585 | 7.0 | 210 | 1.0975 | 0.8981 | 0.8991 | 0.8982 | 6.9497 | 16 | 3 | 11.135 | 2.2086 |
| 0.7086 | 8.0 | 240 | 1.1068 | 0.8989 | 0.8971 | 0.8976 | 6.7374 | 15 | 2 | 10.9411 | 1.1043 |
| 0.6541 | 9.0 | 270 | 1.1340 | 0.898 | 0.902 | 0.8996 | 7.1239 | 17 | 2 | 11.4258 | 2.8221 |
| 0.6167 | 10.0 | 300 | 1.1316 | 0.8981 | 0.8996 | 0.8984 | 6.962 | 16 | 2 | 11.1436 | 2.6994 |
| 0.5817 | 11.0 | 330 | 1.1507 | 0.8984 | 0.8995 | 0.8985 | 6.9264 | 16 | 2 | 11.2466 | 2.3313 |
| 0.547 | 12.0 | 360 | 1.1416 | 0.899 | 0.8993 | 0.8988 | 6.8601 | 15 | 2 | 11.1865 | 1.5951 |
| 0.5181 | 13.0 | 390 | 1.1775 | 0.8987 | 0.8993 | 0.8986 | 6.8969 | 15 | 3 | 11.1571 | 1.8405 |
| 0.4874 | 14.0 | 420 | 1.2131 | 0.8969 | 0.8999 | 0.8979 | 7.0638 | 16 | 3 | 11.3877 | 2.454 |
| 0.461 | 15.0 | 450 | 1.2263 | 0.9009 | 0.9001 | 0.9001 | 6.8871 | 15 | 2 | 11.1497 | 1.1043 |
| 0.4449 | 16.0 | 480 | 1.2215 | 0.898 | 0.8998 | 0.8984 | 7.0184 | 15 | 2 | 11.3509 | 2.2086 |
| 0.4148 | 17.0 | 510 | 1.2528 | 0.8991 | 0.8997 | 0.899 | 6.8748 | 15 | 2 | 11.1939 | 1.227 |
| 0.3967 | 18.0 | 540 | 1.2512 | 0.8989 | 0.9006 | 0.8994 | 6.9853 | 15 | 2 | 11.3227 | 1.3497 |
| 0.385 | 19.0 | 570 | 1.2683 | 0.898 | 0.8992 | 0.8982 | 6.9755 | 17 | 2 | 11.3227 | 2.454 |
| 0.3667 | 20.0 | 600 | 1.2804 | 0.898 | 0.8982 | 0.8976 | 6.8528 | 15 | 2 | 11.1337 | 1.7178 |
| 0.3579 | 21.0 | 630 | 1.2892 | 0.9006 | 0.899 | 0.8993 | 6.762 | 15 | 2 | 11.0761 | 1.8405 |
| 0.3358 | 22.0 | 660 | 1.3081 | 0.8983 | 0.8996 | 0.8985 | 6.8957 | 15 | 2 | 11.3166 | 1.5951 |
| 0.3261 | 23.0 | 690 | 1.3189 | 0.8988 | 0.8997 | 0.8988 | 6.9571 | 15 | 2 | 11.2491 | 2.6994 |
| 0.3222 | 24.0 | 720 | 1.3116 | 0.8975 | 0.9 | 0.8983 | 7.027 | 15 | 2 | 11.3926 | 2.9448 |
| 0.2977 | 25.0 | 750 | 1.3295 | 0.8988 | 0.9005 | 0.8992 | 6.9755 | 15 | 2 | 11.308 | 1.9632 |
| 0.2977 | 26.0 | 780 | 1.3368 | 0.8972 | 0.9009 | 0.8986 | 7.0442 | 15 | 2 | 11.4184 | 2.5767 |
| 0.2795 | 27.0 | 810 | 1.3515 | 0.9015 | 0.9025 | 0.9015 | 6.9117 | 15 | 2 | 11.2601 | 2.2086 |
| 0.2758 | 28.0 | 840 | 1.3645 | 0.9018 | 0.9013 | 0.9011 | 6.8466 | 15 | 2 | 11.1264 | 1.9632 |
| 0.2696 | 29.0 | 870 | 1.3675 | 0.899 | 0.9002 | 0.8991 | 6.9399 | 15 | 2 | 11.2687 | 1.9632 |
| 0.261 | 30.0 | 900 | 1.3896 | 0.8986 | 0.901 | 0.8993 | 7.0282 | 17 | 2 | 11.3902 | 2.0859 |
| 0.2632 | 31.0 | 930 | 1.3802 | 0.8971 | 0.8988 | 0.8975 | 6.9546 | 15 | 2 | 11.2528 | 2.3313 |
| 0.2488 | 32.0 | 960 | 1.3919 | 0.8998 | 0.9011 | 0.9 | 6.9656 | 15 | 2 | 11.2748 | 2.454 |
| 0.2467 | 33.0 | 990 | 1.3973 | 0.8996 | 0.9013 | 0.9 | 6.9755 | 15 | 2 | 11.3031 | 2.2086 |
| 0.2384 | 34.0 | 1020 | 1.4041 | 0.8988 | 0.9001 | 0.899 | 6.9865 | 15 | 2 | 11.2785 | 2.3313 |
| 0.2369 | 35.0 | 1050 | 1.4044 | 0.8986 | 0.9005 | 0.8991 | 6.9595 | 15 | 2 | 11.2785 | 1.9632 |
| 0.2307 | 36.0 | 1080 | 1.4109 | 0.8985 | 0.9 | 0.8988 | 6.9436 | 15 | 2 | 11.2712 | 2.0859 |
| 0.2285 | 37.0 | 1110 | 1.4118 | 0.8987 | 0.8995 | 0.8987 | 6.9067 | 15 | 2 | 11.2037 | 1.7178 |
| 0.2294 | 38.0 | 1140 | 1.4171 | 0.8992 | 0.9007 | 0.8995 | 6.9436 | 15 | 2 | 11.2712 | 1.9632 |
| 0.2294 | 39.0 | 1170 | 1.4188 | 0.8988 | 0.9008 | 0.8994 | 6.9656 | 15 | 2 | 11.3006 | 2.2086 |
| 0.2261 | 40.0 | 1200 | 1.4183 | 0.8989 | 0.9008 | 0.8994 | 6.9571 | 15 | 2 | 11.2896 | 2.0859 |
### Framework versions
- Transformers 4.33.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.13.3
|
pepoo20/whisper_lora_big-LORA-colab
|
pepoo20
| 2023-10-06T09:19:20Z | 2 | 0 |
peft
|
[
"peft",
"arxiv:1910.09700",
"base_model:openai/whisper-large-v2",
"base_model:adapter:openai/whisper-large-v2",
"region:us"
] | null | 2023-10-06T09:19:18Z |
---
library_name: peft
base_model: openai/whisper-large-v2
---
# 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]
- **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 Data 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 Data 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]
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: True
- load_in_4bit: False
- 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: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
### Framework versions
- PEFT 0.6.0.dev0
|
sksayril/bpt-v-4-Bengali
|
sksayril
| 2023-10-06T09:17:47Z | 2 | 0 |
peft
|
[
"peft",
"bloom",
"arxiv:1910.09700",
"base_model:bigscience/bloom-560m",
"base_model:adapter:bigscience/bloom-560m",
"region:us"
] | null | 2023-10-06T06:08:47Z |
---
library_name: peft
base_model: bigscience/bloom-560m
---
# 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]
- **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 Data 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 Data 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]
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: True
- load_in_4bit: False
- 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: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
### Framework versions
- PEFT 0.6.0.dev0
|
ldos/text_shortening_model_v74
|
ldos
| 2023-10-06T08:49:53Z | 107 | 0 |
transformers
|
[
"transformers",
"pytorch",
"t5",
"text2text-generation",
"generated_from_trainer",
"base_model:google-t5/t5-small",
"base_model:finetune:google-t5/t5-small",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-10-06T08:40:11Z |
---
license: apache-2.0
base_model: t5-small
tags:
- generated_from_trainer
model-index:
- name: text_shortening_model_v74
results: []
---
<!-- 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. -->
# text_shortening_model_v74
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.2644
- Bert precision: 0.8826
- Bert recall: 0.8851
- Bert f1-score: 0.8832
- Average word count: 6.7137
- Max word count: 16
- Min word count: 2
- Average token count: 10.6547
- % shortened texts with length > 12: 2.6026
## 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: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 40
### Training results
| Training Loss | Epoch | Step | Validation Loss | Bert precision | Bert recall | Bert f1-score | Average word count | Max word count | Min word count | Average token count | % shortened texts with length > 12 |
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:-----------:|:-------------:|:------------------:|:--------------:|:--------------:|:-------------------:|:----------------------------------:|
| 2.5095 | 1.0 | 37 | 1.9952 | 0.8247 | 0.8396 | 0.8308 | 8.5926 | 19 | 0 | 12.961 | 12.8128 |
| 2.1271 | 2.0 | 74 | 1.7552 | 0.8393 | 0.8454 | 0.841 | 7.7247 | 17 | 0 | 11.7738 | 9.3093 |
| 1.9629 | 3.0 | 111 | 1.6420 | 0.8552 | 0.8582 | 0.8556 | 7.2022 | 17 | 1 | 11.3193 | 6.5065 |
| 1.8511 | 4.0 | 148 | 1.5687 | 0.8646 | 0.8639 | 0.8634 | 6.8078 | 17 | 1 | 10.8539 | 4.1041 |
| 1.7806 | 5.0 | 185 | 1.5196 | 0.8684 | 0.8693 | 0.8681 | 6.8278 | 16 | 1 | 10.8438 | 3.8038 |
| 1.7193 | 6.0 | 222 | 1.4840 | 0.8713 | 0.8736 | 0.8717 | 6.8388 | 18 | 2 | 10.8318 | 3.4034 |
| 1.6763 | 7.0 | 259 | 1.4540 | 0.8756 | 0.8765 | 0.8754 | 6.7528 | 18 | 2 | 10.6847 | 3.003 |
| 1.6389 | 8.0 | 296 | 1.4316 | 0.8766 | 0.8785 | 0.8769 | 6.7628 | 16 | 2 | 10.6917 | 2.6026 |
| 1.6146 | 9.0 | 333 | 1.4149 | 0.8771 | 0.8798 | 0.8778 | 6.8018 | 15 | 2 | 10.7177 | 2.7027 |
| 1.597 | 10.0 | 370 | 1.3986 | 0.8782 | 0.8811 | 0.879 | 6.7998 | 15 | 2 | 10.7067 | 2.5025 |
| 1.5761 | 11.0 | 407 | 1.3860 | 0.8792 | 0.8815 | 0.8797 | 6.7588 | 15 | 2 | 10.6496 | 2.3023 |
| 1.5456 | 12.0 | 444 | 1.3747 | 0.8792 | 0.8813 | 0.8797 | 6.7387 | 16 | 2 | 10.6376 | 2.2022 |
| 1.533 | 13.0 | 481 | 1.3647 | 0.88 | 0.8823 | 0.8805 | 6.7347 | 16 | 2 | 10.6276 | 2.1021 |
| 1.5142 | 14.0 | 518 | 1.3536 | 0.8805 | 0.8822 | 0.8808 | 6.7047 | 16 | 2 | 10.5746 | 1.9019 |
| 1.514 | 15.0 | 555 | 1.3429 | 0.8803 | 0.882 | 0.8805 | 6.6847 | 16 | 2 | 10.5606 | 1.7017 |
| 1.4973 | 16.0 | 592 | 1.3353 | 0.8805 | 0.8828 | 0.881 | 6.7467 | 16 | 2 | 10.6627 | 2.1021 |
| 1.4792 | 17.0 | 629 | 1.3277 | 0.8811 | 0.8829 | 0.8814 | 6.7077 | 16 | 2 | 10.6166 | 2.002 |
| 1.4669 | 18.0 | 666 | 1.3206 | 0.8815 | 0.8831 | 0.8817 | 6.6927 | 16 | 2 | 10.6016 | 2.1021 |
| 1.4667 | 19.0 | 703 | 1.3141 | 0.881 | 0.8831 | 0.8815 | 6.7167 | 16 | 2 | 10.6306 | 2.1021 |
| 1.4497 | 20.0 | 740 | 1.3097 | 0.8808 | 0.883 | 0.8813 | 6.7227 | 16 | 2 | 10.6416 | 2.1021 |
| 1.4533 | 21.0 | 777 | 1.3053 | 0.8814 | 0.8831 | 0.8817 | 6.6997 | 16 | 2 | 10.6086 | 2.1021 |
| 1.4408 | 22.0 | 814 | 1.2998 | 0.8808 | 0.8825 | 0.881 | 6.7037 | 16 | 2 | 10.6076 | 2.2022 |
| 1.4343 | 23.0 | 851 | 1.2958 | 0.8807 | 0.8829 | 0.8812 | 6.7297 | 16 | 2 | 10.6306 | 2.3023 |
| 1.4295 | 24.0 | 888 | 1.2926 | 0.881 | 0.8833 | 0.8816 | 6.7427 | 16 | 2 | 10.6486 | 2.4024 |
| 1.4219 | 25.0 | 925 | 1.2887 | 0.8812 | 0.8835 | 0.8818 | 6.7327 | 16 | 2 | 10.6426 | 2.4024 |
| 1.4045 | 26.0 | 962 | 1.2855 | 0.8814 | 0.8836 | 0.8819 | 6.7187 | 16 | 2 | 10.6256 | 2.4024 |
| 1.409 | 27.0 | 999 | 1.2826 | 0.8817 | 0.884 | 0.8823 | 6.7217 | 16 | 2 | 10.6456 | 2.6026 |
| 1.3994 | 28.0 | 1036 | 1.2803 | 0.8826 | 0.8848 | 0.8831 | 6.7047 | 16 | 2 | 10.6226 | 2.7027 |
| 1.3905 | 29.0 | 1073 | 1.2778 | 0.8823 | 0.8847 | 0.8829 | 6.7267 | 16 | 2 | 10.6507 | 2.8028 |
| 1.4014 | 30.0 | 1110 | 1.2751 | 0.8821 | 0.8845 | 0.8827 | 6.7237 | 16 | 2 | 10.6466 | 2.8028 |
| 1.3946 | 31.0 | 1147 | 1.2732 | 0.8826 | 0.8849 | 0.8831 | 6.7167 | 16 | 2 | 10.6426 | 2.8028 |
| 1.3915 | 32.0 | 1184 | 1.2712 | 0.8823 | 0.8845 | 0.8828 | 6.7057 | 16 | 2 | 10.6336 | 2.7027 |
| 1.3904 | 33.0 | 1221 | 1.2695 | 0.8824 | 0.8847 | 0.883 | 6.7047 | 16 | 2 | 10.6376 | 2.7027 |
| 1.3843 | 34.0 | 1258 | 1.2684 | 0.8828 | 0.885 | 0.8833 | 6.7097 | 16 | 2 | 10.6406 | 2.6026 |
| 1.3875 | 35.0 | 1295 | 1.2672 | 0.8827 | 0.8852 | 0.8834 | 6.7217 | 16 | 2 | 10.6607 | 2.6026 |
| 1.3794 | 36.0 | 1332 | 1.2661 | 0.8828 | 0.8851 | 0.8834 | 6.7087 | 16 | 2 | 10.6426 | 2.6026 |
| 1.3906 | 37.0 | 1369 | 1.2654 | 0.8828 | 0.8853 | 0.8835 | 6.7177 | 16 | 2 | 10.6567 | 2.6026 |
| 1.3841 | 38.0 | 1406 | 1.2648 | 0.8826 | 0.8851 | 0.8833 | 6.7107 | 16 | 2 | 10.6476 | 2.6026 |
| 1.3761 | 39.0 | 1443 | 1.2645 | 0.8825 | 0.885 | 0.8832 | 6.7137 | 16 | 2 | 10.6537 | 2.6026 |
| 1.3797 | 40.0 | 1480 | 1.2644 | 0.8826 | 0.8851 | 0.8832 | 6.7137 | 16 | 2 | 10.6547 | 2.6026 |
### Framework versions
- Transformers 4.33.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.13.3
|
stabilityai/japanese-stablelm-instruct-alpha-7b-v2
|
stabilityai
| 2023-10-06T08:40:24Z | 292 | 22 |
transformers
|
[
"transformers",
"safetensors",
"text-generation",
"japanese-stablelm",
"causal-lm",
"custom_code",
"ja",
"license:apache-2.0",
"autotrain_compatible",
"region:us"
] |
text-generation
| 2023-09-26T05:28:31Z |
---
language:
- ja
tags:
- japanese-stablelm
- causal-lm
pipeline_tag: text-generation
license: apache-2.0
---
# Japanese-StableLM-Instruct-Alpha-7B-v2
> "A parrot able to speak Japanese, ukiyoe, edo period" — [Stable Diffusion XL](https://clipdrop.co/stable-diffusion)
## Model Description
`japanese-stablelm-instruct-alpha-7b-v2` is a 7B parameter decoder-only language models pre-trained built on top of the [`Japanese-StableLM-Base-Alpha-7B`](https://huggingface.co/stabilityai/japanese-stablelm-base-alpha-7b) model and further fine-tuned on various instruction-following datasets.
## Usage
First install additional dependencies in [requirements.txt](https://huggingface.co/stabilityai/japanese-stablelm-base-alpha-7b/blob/main/requirements.txt):
```sh
pip install sentencepiece einops
```
Then start generating text with `japanese-stablelm-instruct-alpha-7b-v2` by using the following code snippet:
```python
import torch
from transformers import LlamaTokenizer, AutoModelForCausalLM
tokenizer = LlamaTokenizer.from_pretrained(
"novelai/nerdstash-tokenizer-v1", additional_special_tokens=["▁▁"]
)
model = AutoModelForCausalLM.from_pretrained(
"stabilityai/japanese-stablelm-instruct-alpha-7b-v2",
trust_remote_code=True,
torch_dtype=torch.float16,
variant="fp16",
)
model.eval()
if torch.cuda.is_available():
model = model.to("cuda")
def build_prompt(user_query, inputs="", sep="\n\n### "):
sys_msg = "以下は、タスクを説明する指示と、文脈のある入力の組み合わせです。要求を適切に満たす応答を書きなさい。"
p = sys_msg
roles = ["指示", "応答"]
msgs = [": \n" + user_query, ": \n"]
if inputs:
roles.insert(1, "入力")
msgs.insert(1, ": \n" + inputs)
for role, msg in zip(roles, msgs):
p += sep + role + msg
return p
# Infer with prompt without any additional input
user_inputs = {
"user_query": "与えられたことわざの意味を小学生でも分かるように教えてください。",
"inputs": "情けは人のためならず"
}
prompt = build_prompt(**user_inputs)
input_ids = tokenizer.encode(
prompt,
add_special_tokens=False,
return_tensors="pt"
)
tokens = model.generate(
input_ids.to(device=model.device),
max_new_tokens=256,
temperature=1,
top_p=0.95,
do_sample=True,
)
out = tokenizer.decode(tokens[0][input_ids.shape[1]:], skip_special_tokens=True).strip()
print(out)
"""
「情けは人のためならず」は、「情けをかけるとその人のためにならない」という意味ではありません。
このことわざは、もともと「誰かのために行動するとその行動が回り回って自分に返ってくる」ということを説いたことわざです。
"""
```
## Model Details
* **Model type**: `japanese-stablelm-instruct-alpha-7b-v2` is an auto-regressive language model based on the NeoX transformer architecture.
* **Language(s)**: Japanese
* **Library**: [GPT-NeoX](https://github.com/EleutherAI/gpt-neox)
* **License**: This model is licensed under [Apache License, Version 2.0](https://www.apache.org/licenses/LICENSE-2.0).
## Training
| Parameters | Hidden Size | Layers | Heads | Sequence Length |
|------------|-------------|--------|-------|-----------------|
| 7B | 4096 | 32 | 32 | 1024 |
### Training Dataset
`japanese-stablelm-instruct-alpha-7b-v2` is fine-tuned on a combination of following datasets:
- [Japanese translation of the Databricks Dolly-15k dataset](https://huggingface.co/datasets/kunishou/databricks-dolly-15k-ja)
- [Japanese translation of the subset of the Anthropic HH dataset](https://huggingface.co/datasets/fujiki/japanese_hh-rlhf-49k)
- [Wikinews](https://ja.wikinews.org/wi) [subset](https://huggingface.co/datasets/fujiki/llm-japanese-dataset_wikinews) of the [izumi-lab/llm-japanese-dataset](https://huggingface.co/datasets/izumi-lab/llm-japanese-dataset)
## Use and Limitations
### Intended Use
This model is intended to be used by the open-source community in chat-like applications in adherence with [Apache-2.0 license](https://www.apache.org/licenses/LICENSE-2.0).
### Limitations and bias
Although the aforementioned datasets help to steer the base language models into "safer" distributions of text, not all biases and toxicity can be mitigated through fine-tuning. We ask that users be mindful of such potential issues that can arise in generated responses. Do not treat model outputs as substitutes for human judgment or as sources of truth. Please use responsibly.
## Authors
- [Meng Lee](https://huggingface.co/leemeng)
- [Fujiki Nakamura](https://huggingface.co/fujiki)
- [Makoto Shing](https://huggingface.co/mkshing)
- [Paul McCann](https://huggingface.co/polm-stability)
- [Takuya Akiba](https://huggingface.co/iwiwi)
- [Naoki Orii](https://huggingface.co/mrorii)
## Acknowledgements
We are utilizing the v1 version of the [novelai-tokenizer](https://github.com/NovelAI/novelai-tokenizer), introduced by [NovelAI](https://novelai.net/), because it processes both Japanese and English text both effectively and efficiently. We extend our gratitude to NovelAI for allowing us to use their remarkable work. For more details about the tokenizer, please refer to their [blog post](https://blog.novelai.net/novelais-new-llm-tokenizer-5bc140e17642).
We are grateful for the contributions of the EleutherAI Polyglot-JA team in helping us to collect a large amount of pre-training data in Japanese. Polyglot-JA members includes Hyunwoong Ko (Project Lead), Fujiki Nakamura (originally started this project when he committed to the Polyglot team), Yunho Mo, Minji Jung, KeunSeok Im, and Su-Kyeong Jang.
We are also appreciative of [AI Novelist/Sta (Bit192, Inc.)](https://ai-novel.com/index.php) and the numerous contributors from [Stable Community Japan](https://discord.gg/VPrcE475HB) for assisting us in gathering a large amount of high-quality Japanese textual data for model training.
## How to cite
```bibtext
@misc{JapaneseStableLMInstructAlpha7Bv2,
url={[https://huggingface.co/stabilityai/japanese-stablelm-instruct-alpha-7b-v2](https://huggingface.co/stabilityai/japanese-stablelm-instruct-alpha-7b-v2)},
title={Japanese StableLM Instruct Alpha 7B v2},
author={Lee, Meng and Nakamura, Fujiki and Shing, Makoto and McCann, Paul and Akiba, Takuya and Orii, Naoki}
}
```
## Citations
```bibtex
@misc{alpaca,
author = {Rohan Taori and Ishaan Gulrajani and Tianyi Zhang and Yann Dubois and Xuechen Li and Carlos Guestrin and Percy Liang and Tatsunori B. Hashimoto },
title = {Stanford Alpaca: An Instruction-following LLaMA model},
year = {2023},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {\url{https://github.com/tatsu-lab/stanford_alpaca}},
}
```
```bibtext
@software{gpt-neox-library,
title = {{GPT-NeoX: Large Scale Autoregressive Language Modeling in PyTorch}},
author = {Andonian, Alex and Anthony, Quentin and Biderman, Stella and Black, Sid and Gali, Preetham and Gao, Leo and Hallahan, Eric and Levy-Kramer, Josh and Leahy, Connor and Nestler, Lucas and Parker, Kip and Pieler, Michael and Purohit, Shivanshu and Songz, Tri and Phil, Wang and Weinbach, Samuel},
url = {https://www.github.com/eleutherai/gpt-neox},
doi = {10.5281/zenodo.5879544},
month = {8},
year = {2021},
version = {0.0.1},
}
```
|
posicube/Llama2-chat-AYB-13B
|
posicube
| 2023-10-06T08:40:15Z | 1,406 | 14 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"arxiv:2306.02707",
"license:llama2",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-10-03T03:06:14Z |
---
license: llama2
---
This is a model diverged from Llama-2-13b-chat-hf. We hypotheize that if we find a method to ensemble the top rankers in each benchmark effectively, its performance maximizes as well. Following this intuition, we ensembled the top models in each benchmarks to create our model.
# Model Details
- **Developed by**: Posicube Inc.
- **Backbone Model**: LLaMA-2-13b-chat
- **Library**: HuggingFace Transformers
- **Used Dataset Details**
Orca-style datasets, Alpaca-style datasets
# Evaluation
We achieved the top ranker among 13B models with this model at Oct-3rd 2023.
| Metric |Scores on Leaderboard| Our results |
|---------------------|---------------------|-------------|
| ARC (25-shot) | 63.4 | 63.48 |
| HellaSwag (10-shot) | 84.79 | 84.87 |
| MMLU (5-shot) | 59.34 | 59.59 |
| TruthfulQA (0-shot) | 55.62 | 55.22 |
| Avg. | 65.79 | 65.78 |
# Limitations & Biases:
Llama2 and fine-tuned variants are a new technology that carries risks with 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 2 and any fine-tuned varient'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 2 variants, developers should perform safety testing and tuning tailored to their specific applications of the model.
Please see the Responsible Use Guide available at https://ai.meta.com/llama/responsible-use-guide/
# License Disclaimer:
This model is bound by the license & usage restrictions of the original Llama-2 model. And comes with no warranty or gurantees of any kind.
# Contact Us
[Posicube](https://www.posicube.com/)
# Citiation:
Please kindly cite using the following BibTeX:
```bibtex
@misc{mukherjee2023orca,
title={Orca: Progressive Learning from Complex Explanation Traces of GPT-4},
author={Subhabrata Mukherjee and Arindam Mitra and Ganesh Jawahar and Sahaj Agarwal and Hamid Palangi and Ahmed Awadallah},
year={2023},
eprint={2306.02707},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
```bibtex
@software{touvron2023llama2,
title={Llama 2: Open Foundation and Fine-Tuned Chat Models},
author={Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava,
Shruti Bhosale, Dan Bikel, Lukas Blecher, Cristian Canton Ferrer, Moya Chen, Guillem Cucurull, David Esiobu, Jude Fernandes, Jeremy Fu, Wenyin Fu, Brian Fuller,
Cynthia Gao, Vedanuj Goswami, Naman Goyal, Anthony Hartshorn, Saghar Hosseini, Rui Hou, Hakan Inan, Marcin Kardas, Viktor Kerkez Madian Khabsa, Isabel Kloumann,
Artem Korenev, Punit Singh Koura, Marie-Anne Lachaux, Thibaut Lavril, Jenya Lee, Diana Liskovich, Yinghai Lu, Yuning Mao, Xavier Martinet, Todor Mihaylov,
Pushkar Mishra, Igor Molybog, Yixin Nie, Andrew Poulton, Jeremy Reizenstein, Rashi Rungta, Kalyan Saladi, Alan Schelten, Ruan Silva, Eric Michael Smith,
Ranjan Subramanian, Xiaoqing Ellen Tan, Binh Tang, Ross Taylor, Adina Williams, Jian Xiang Kuan, Puxin Xu , Zheng Yan, Iliyan Zarov, Yuchen Zhang, Angela Fan,
Melanie Kambadur, Sharan Narang, Aurelien Rodriguez, Robert Stojnic, Sergey Edunov, Thomas Scialom},
year={2023}
}
```
|
ldos/text_shortening_model_v73
|
ldos
| 2023-10-06T08:39:06Z | 103 | 0 |
transformers
|
[
"transformers",
"pytorch",
"t5",
"text2text-generation",
"generated_from_trainer",
"base_model:google-t5/t5-small",
"base_model:finetune:google-t5/t5-small",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-10-06T08:29:43Z |
---
license: apache-2.0
base_model: t5-small
tags:
- generated_from_trainer
model-index:
- name: text_shortening_model_v73
results: []
---
<!-- 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. -->
# text_shortening_model_v73
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6126
- Bert precision: 0.9015
- Bert recall: 0.9014
- Bert f1-score: 0.901
- Average word count: 6.4004
- Max word count: 16
- Min word count: 2
- Average token count: 10.4705
- % shortened texts with length > 12: 1.1011
## 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: 0.0005
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 40
### Training results
| Training Loss | Epoch | Step | Validation Loss | Bert precision | Bert recall | Bert f1-score | Average word count | Max word count | Min word count | Average token count | % shortened texts with length > 12 |
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:-----------:|:-------------:|:------------------:|:--------------:|:--------------:|:-------------------:|:----------------------------------:|
| 1.5857 | 1.0 | 37 | 1.1932 | 0.8846 | 0.8848 | 0.8842 | 6.5315 | 16 | 1 | 10.4525 | 1.7017 |
| 1.184 | 2.0 | 74 | 1.0965 | 0.8918 | 0.8915 | 0.8911 | 6.4855 | 17 | 2 | 10.4735 | 0.5005 |
| 1.0114 | 3.0 | 111 | 1.0773 | 0.8895 | 0.8962 | 0.8924 | 6.8959 | 18 | 2 | 10.995 | 1.3013 |
| 0.8887 | 4.0 | 148 | 1.0798 | 0.8947 | 0.8936 | 0.8937 | 6.4454 | 17 | 2 | 10.4605 | 1.8018 |
| 0.7851 | 5.0 | 185 | 1.0807 | 0.8941 | 0.8948 | 0.894 | 6.5676 | 16 | 2 | 10.6016 | 1.6016 |
| 0.7116 | 6.0 | 222 | 1.1002 | 0.8984 | 0.8978 | 0.8976 | 6.4605 | 15 | 2 | 10.4174 | 1.2012 |
| 0.6472 | 7.0 | 259 | 1.1171 | 0.8982 | 0.8997 | 0.8985 | 6.5836 | 16 | 2 | 10.6426 | 1.3013 |
| 0.5872 | 8.0 | 296 | 1.1196 | 0.8998 | 0.9015 | 0.9002 | 6.5415 | 16 | 2 | 10.6226 | 1.5015 |
| 0.5393 | 9.0 | 333 | 1.1739 | 0.9007 | 0.8979 | 0.8988 | 6.3333 | 16 | 2 | 10.3063 | 1.1011 |
| 0.4879 | 10.0 | 370 | 1.2079 | 0.8997 | 0.8983 | 0.8985 | 6.3343 | 15 | 2 | 10.2913 | 1.001 |
| 0.4615 | 11.0 | 407 | 1.2230 | 0.8988 | 0.8997 | 0.8988 | 6.5165 | 15 | 2 | 10.6426 | 1.3013 |
| 0.4245 | 12.0 | 444 | 1.2325 | 0.8996 | 0.8979 | 0.8983 | 6.3704 | 15 | 2 | 10.4334 | 1.3013 |
| 0.3973 | 13.0 | 481 | 1.2657 | 0.8973 | 0.8987 | 0.8975 | 6.4855 | 15 | 2 | 10.5876 | 1.6016 |
| 0.3658 | 14.0 | 518 | 1.2875 | 0.8985 | 0.8993 | 0.8984 | 6.4735 | 15 | 2 | 10.5355 | 1.2012 |
| 0.3422 | 15.0 | 555 | 1.3202 | 0.9002 | 0.8991 | 0.8992 | 6.2873 | 14 | 2 | 10.3594 | 1.001 |
| 0.3271 | 16.0 | 592 | 1.3315 | 0.9006 | 0.9 | 0.8998 | 6.3784 | 15 | 2 | 10.4454 | 0.9009 |
| 0.305 | 17.0 | 629 | 1.3441 | 0.8994 | 0.9005 | 0.8995 | 6.4705 | 16 | 2 | 10.5906 | 1.2012 |
| 0.2847 | 18.0 | 666 | 1.3648 | 0.8997 | 0.8989 | 0.8989 | 6.3584 | 14 | 2 | 10.4244 | 0.9009 |
| 0.2707 | 19.0 | 703 | 1.3837 | 0.9005 | 0.9011 | 0.9003 | 6.4545 | 16 | 2 | 10.5365 | 1.3013 |
| 0.254 | 20.0 | 740 | 1.4180 | 0.8997 | 0.9006 | 0.8997 | 6.4444 | 15 | 2 | 10.5516 | 1.2012 |
| 0.2421 | 21.0 | 777 | 1.4100 | 0.9014 | 0.903 | 0.9017 | 6.4755 | 16 | 2 | 10.6016 | 0.9009 |
| 0.2301 | 22.0 | 814 | 1.4437 | 0.9 | 0.901 | 0.9 | 6.4825 | 15 | 2 | 10.5626 | 0.8008 |
| 0.2183 | 23.0 | 851 | 1.4762 | 0.9003 | 0.9014 | 0.9004 | 6.4995 | 16 | 2 | 10.6116 | 1.3013 |
| 0.2148 | 24.0 | 888 | 1.4815 | 0.9007 | 0.9014 | 0.9006 | 6.4484 | 16 | 2 | 10.5495 | 1.1011 |
| 0.2013 | 25.0 | 925 | 1.5039 | 0.9018 | 0.9015 | 0.9012 | 6.4144 | 15 | 2 | 10.4925 | 1.001 |
| 0.1924 | 26.0 | 962 | 1.5217 | 0.9013 | 0.9014 | 0.9009 | 6.4024 | 16 | 2 | 10.4765 | 1.2012 |
| 0.1854 | 27.0 | 999 | 1.5125 | 0.902 | 0.9014 | 0.9012 | 6.3774 | 16 | 2 | 10.4565 | 1.1011 |
| 0.1769 | 28.0 | 1036 | 1.5384 | 0.8998 | 0.9011 | 0.9 | 6.4925 | 16 | 2 | 10.6106 | 1.001 |
| 0.1713 | 29.0 | 1073 | 1.5627 | 0.9012 | 0.9018 | 0.901 | 6.4715 | 16 | 2 | 10.5395 | 1.2012 |
| 0.1685 | 30.0 | 1110 | 1.5473 | 0.9011 | 0.9004 | 0.9002 | 6.4064 | 16 | 2 | 10.4484 | 1.1011 |
| 0.1681 | 31.0 | 1147 | 1.5592 | 0.9018 | 0.9018 | 0.9013 | 6.4194 | 15 | 2 | 10.5165 | 0.8008 |
| 0.1599 | 32.0 | 1184 | 1.5800 | 0.9006 | 0.9007 | 0.9002 | 6.4254 | 16 | 2 | 10.5005 | 1.001 |
| 0.1509 | 33.0 | 1221 | 1.5822 | 0.9012 | 0.9005 | 0.9004 | 6.3994 | 16 | 2 | 10.4314 | 1.001 |
| 0.1509 | 34.0 | 1258 | 1.5924 | 0.9013 | 0.9008 | 0.9006 | 6.4084 | 16 | 2 | 10.4655 | 1.1011 |
| 0.1408 | 35.0 | 1295 | 1.6045 | 0.9028 | 0.9024 | 0.9021 | 6.4074 | 16 | 2 | 10.4845 | 1.2012 |
| 0.1487 | 36.0 | 1332 | 1.6133 | 0.9014 | 0.9012 | 0.9008 | 6.4244 | 16 | 2 | 10.4775 | 1.001 |
| 0.1444 | 37.0 | 1369 | 1.6157 | 0.9016 | 0.9016 | 0.9012 | 6.4304 | 16 | 2 | 10.5045 | 1.2012 |
| 0.1418 | 38.0 | 1406 | 1.6105 | 0.9012 | 0.9011 | 0.9006 | 6.4084 | 16 | 2 | 10.4615 | 1.1011 |
| 0.1402 | 39.0 | 1443 | 1.6116 | 0.9017 | 0.9015 | 0.9011 | 6.3894 | 16 | 2 | 10.4494 | 1.1011 |
| 0.1375 | 40.0 | 1480 | 1.6126 | 0.9015 | 0.9014 | 0.901 | 6.4004 | 16 | 2 | 10.4705 | 1.1011 |
### Framework versions
- Transformers 4.33.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.13.3
|
leoliu-hkustgz/Reinforce-CartPole
|
leoliu-hkustgz
| 2023-10-06T08:32:14Z | 0 | 0 | null |
[
"CartPole-v1",
"reinforce",
"reinforcement-learning",
"custom-implementation",
"deep-rl-class",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-10-06T08:32:05Z |
---
tags:
- CartPole-v1
- reinforce
- reinforcement-learning
- custom-implementation
- deep-rl-class
model-index:
- name: Reinforce-CartPole
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
---
# **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
|
0xC4LL3/DQN_SpaceInvadersNoFrameskip-V4
|
0xC4LL3
| 2023-10-06T08:28:58Z | 0 | 0 |
stable-baselines3
|
[
"stable-baselines3",
"SpaceInvadersNoFrameskip-v4",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-10-06T08:28:18Z |
---
library_name: stable-baselines3
tags:
- SpaceInvadersNoFrameskip-v4
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: DQN
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: SpaceInvadersNoFrameskip-v4
type: SpaceInvadersNoFrameskip-v4
metrics:
- type: mean_reward
value: 714.00 +/- 196.36
name: mean_reward
verified: false
---
# **DQN** Agent playing **SpaceInvadersNoFrameskip-v4**
This is a trained model of a **DQN** agent playing **SpaceInvadersNoFrameskip-v4**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3)
and the [RL Zoo](https://github.com/DLR-RM/rl-baselines3-zoo).
The RL Zoo is a training framework for Stable Baselines3
reinforcement learning agents,
with hyperparameter optimization and pre-trained agents included.
## Usage (with SB3 RL Zoo)
RL Zoo: https://github.com/DLR-RM/rl-baselines3-zoo<br/>
SB3: https://github.com/DLR-RM/stable-baselines3<br/>
SB3 Contrib: https://github.com/Stable-Baselines-Team/stable-baselines3-contrib
Install the RL Zoo (with SB3 and SB3-Contrib):
```bash
pip install rl_zoo3
```
```
# Download model and save it into the logs/ folder
python -m rl_zoo3.load_from_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -orga 0xC4LL3 -f logs/
python -m rl_zoo3.enjoy --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
```
If you installed the RL Zoo3 via pip (`pip install rl_zoo3`), from anywhere you can do:
```
python -m rl_zoo3.load_from_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -orga 0xC4LL3 -f logs/
python -m rl_zoo3.enjoy --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
```
## Training (with the RL Zoo)
```
python -m rl_zoo3.train --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
# Upload the model and generate video (when possible)
python -m rl_zoo3.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/ -orga 0xC4LL3
```
## Hyperparameters
```python
OrderedDict([('batch_size', 32),
('buffer_size', 100000),
('env_wrapper',
['stable_baselines3.common.atari_wrappers.AtariWrapper']),
('exploration_final_eps', 0.01),
('exploration_fraction', 0.1),
('frame_stack', 4),
('gradient_steps', 1),
('learning_rate', 0.0001),
('learning_starts', 100000),
('n_timesteps', 1000000.0),
('optimize_memory_usage', False),
('policy', 'CnnPolicy'),
('target_update_interval', 1000),
('train_freq', 4),
('normalize', False)])
```
# Environment Arguments
```python
{'render_mode': 'rgb_array'}
```
|
mostafaashahin/wav2vec2-large-robust-libri-voiced-v3
|
mostafaashahin
| 2023-10-06T08:04:36Z | 78 | 0 |
transformers
|
[
"transformers",
"pytorch",
"wav2vec2",
"automatic-speech-recognition",
"endpoints_compatible",
"region:us"
] |
automatic-speech-recognition
| 2023-10-06T07:57:16Z |
---
tags:
- automatic-speech-recognition
---
|
nomsgadded/CartPole-v1
|
nomsgadded
| 2023-10-06T08:02:13Z | 0 | 0 | null |
[
"CartPole-v1",
"reinforce",
"reinforcement-learning",
"custom-implementation",
"deep-rl-class",
"model-index",
"region:us"
] |
reinforcement-learning
| 2023-10-06T08:02:06Z |
---
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: 477.64 +/- 11.32
name: mean_reward
verified: false
---
# **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
|
s3nh/MathLLM-MathCoder-L-7B-GGUF
|
s3nh
| 2023-10-06T07:59:04Z | 1 | 0 |
transformers
|
[
"transformers",
"gguf",
"text-generation",
"zh",
"en",
"license:openrail",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-10-06T07:53:10Z |
---
license: openrail
pipeline_tag: text-generation
library_name: transformers
language:
- zh
- en
---
## Original model card
Buy me a coffee if you like this project ;)
<a href="https://www.buymeacoffee.com/s3nh"><img src="https://www.buymeacoffee.com/assets/img/guidelines/download-assets-sm-1.svg" alt=""></a>
#### Description
GGUF Format model files for [This project](https://huggingface.co/MathLLM/MathCoder-L-7B).
### GGUF Specs
GGUF is a format based on the existing GGJT, but makes a few changes to the format to make it more extensible and easier to use. The following features are desired:
Single-file deployment: they can be easily distributed and loaded, and do not require any external files for additional information.
Extensible: new features can be added to GGML-based executors/new information can be added to GGUF models without breaking compatibility with existing models.
mmap compatibility: models can be loaded using mmap for fast loading and saving.
Easy to use: models can be easily loaded and saved using a small amount of code, with no need for external libraries, regardless of the language used.
Full information: all information needed to load a model is contained in the model file, and no additional information needs to be provided by the user.
The key difference between GGJT and GGUF is the use of a key-value structure for the hyperparameters (now referred to as metadata), rather than a list of untyped values.
This allows for new metadata to be added without breaking compatibility with existing models, and to annotate the model with additional information that may be useful for
inference or for identifying the model.
### Perplexity params
Model Measure Q2_K Q3_K_S Q3_K_M Q3_K_L Q4_0 Q4_1 Q4_K_S Q4_K_M Q5_0 Q5_1 Q5_K_S Q5_K_M Q6_K Q8_0 F16
7B perplexity 6.7764 6.4571 6.1503 6.0869 6.1565 6.0912 6.0215 5.9601 5.9862 5.9481 5.9419 5.9208 5.9110 5.9070 5.9066
13B perplexity 5.8545 5.6033 5.4498 5.4063 5.3860 5.3608 5.3404 5.3002 5.2856 5.2706 5.2785 5.2638 5.2568 5.2548 5.2543
### inference
TODO
# Original model card
|
tyasnk/paraphrase-multilingual-MiniLM-L12-v2-id
|
tyasnk
| 2023-10-06T07:45:47Z | 11 | 0 |
sentence-transformers
|
[
"sentence-transformers",
"pytorch",
"bert",
"feature-extraction",
"sentence-similarity",
"transformers",
"autotrain_compatible",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] |
sentence-similarity
| 2023-10-06T07:02:56Z |
---
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
---
|
Vishal24/adapters-v5
|
Vishal24
| 2023-10-06T07:45:20Z | 1 | 0 |
peft
|
[
"peft",
"arxiv:1910.09700",
"base_model:meta-llama/Llama-2-7b-chat-hf",
"base_model:adapter:meta-llama/Llama-2-7b-chat-hf",
"region:us"
] | null | 2023-10-06T07:45:12Z |
---
library_name: peft
base_model: meta-llama/Llama-2-7b-chat-hf
---
# 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]
- **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 Data 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 Data 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]
## 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 0.6.0.dev0
## 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 0.6.0.dev0
|
Brazilia/Phuong
|
Brazilia
| 2023-10-06T07:37:20Z | 0 | 0 | null |
[
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-10-06T07:34:11Z |
---
license: creativeml-openrail-m
---
|
juyongjiang/CodeUp-Llama-2-13b-chat-hf
|
juyongjiang
| 2023-10-06T06:59:02Z | 57 | 32 |
transformers
|
[
"transformers",
"pytorch",
"llama",
"text-generation",
"text-to-code",
"multilingual-code-generation",
"en",
"arxiv:2106.09685",
"license:openrail++",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-08-01T05:44:51Z |
---
license: openrail++
language:
- en
tags:
- text-to-code
- multilingual-code-generation
---
<!-- <p align="center" width="70%">
<img src="assets/Logo.jpg" alt="HKUST CodeUp" style="width: 50%; min-width: 250px; display: block; margin: auto;">
</p> -->
# CodeUp: A Multilingual Code Generation Llama2 Model with Parameter-Efficient Instruction-Tuning on a Single RTX 3090
## Description
In recent years, large language models (LLMs) have shown exceptional capabilities in a wide range of applications due to their fantastic emergence ability. To align with human preference, instruction-tuning and reinforcement learning from human feedback (RLHF) are proposed for Chat-based LLMs (e.g., ChatGPT, GPT-4). However, these LLMs (except for Codex) primarily focus on the general domain and are not specifically designed for the code domain. Although Codex provides an alternative choice, it is a closed-source model developed by OpenAI. Hence, it is imperative to develop open-source instruction-following LLMs for the code domain.
However, the large-scale number of LLMs' parameters ($\ge$7B) and training datasets require a vast amount of computational resources, which significantly impedes the development of training and inference on consumer hardware.
To handle these challenges, in this project, we adopt the latest powerful foundation model `Llama 2` and construct high-quality instruction-following data for code generation tasks, and propose an instruction-following multilingual code generation Llama2 model. Meanwhile, to make it fit an academic budget and consumer hardware (e.g., a single RTX 3090) based on `Alpaca-LoRA`, we equip `CodeUp` with the advanced parameter-efficient fine-tuning (PEFT) methods (e.g., [LoRA](https://arxiv.org/abs/2106.09685)) which enable efficient adaptation of pre-trained language models (PLMs, also known as foundation model) to various downstream applications without fine-tuning the entire model's parameters. The overall training recipe is as follows.
## NL2Code Data Release
Recently, it has attracted significant attention to exploiting much larger and more powerful LLMs (e.g., ChatGPT, GPT-4) to self-generate instruction-following data by delicate prompt design. However, many approaches primarily focus on the general domain and lack code-specific domain considerations. To this end, [Code Alpaca](https://github.com/sahil280114/codealpaca) follows the previous Self-Instruct paper [3] and [Stanford Alpaca repo](https://github.com/tatsu-lab/stanford_alpaca) with some code-related modifications to conduct 20K instruction-following data `data/code_alpaca_20k.json` for code generation tasks. This `JSON` file following `alpaca_data.json` format is a list of dictionaries; each dictionary contains the following fields:
- `instruction`: `str`, describes the task the model should perform. Each of the 20K instructions is unique.
- `input`: `str`, optional context or input for the task. For example, when the instruction is "Amend the following SQL query to select distinct elements", the input is the SQL query. Around 40% of the examples have an input.
- `output`: `str`, the answer to the instruction as generated by `text-davinci-003`.
### High-quality Data Filter
However, after carefully checking the LLMs-self-generated data, we observe three critical problems that may hinder LLMs' instruction learning due to ambiguous and irrelevant noise. That is
1. When `instruction` doesn't specify the programming language (PL) of implementation, the `output` appears with diverse options, e.g., Python, C++, and JavaScript.
2. It is ambiguous to identify which programming language `output` is implemented by.
3. Both `instruction` and `output` are irrelevant to the code-specific domain.
Hence, we filter the ambiguous and irrelevant data by rigorous design to obtain high-quality instruction data. Specifically, to solve 1) we set Python as the default PL of implementation and use [Guesslang](https://guesslang.readthedocs.io/en/latest/) package to detect the PL of a given source code in `output`. If the Python is detected, this prompt is retained. Otherwise, it will be filtered. 2) and 3) In these cases, we delete these prompts. After that, about 5K low-quality instruction data is filtered. To supplement the high-quality instruction data, we further integrate the `data/new_codealpaca.json` data (about 4.5K) under the above filter rules.
This way, we gain the 19K high-quality instruction data of code generation. The following is the instruction number distribution of each PL with Radar visualization before and after filtering.
<!-- | Raw Data (20K + 4K)| Filtered Data (19K) |
| -- | -- |
| <center><img src="assets/PL_Raw.png" width="100%"></center> | <center><img src="assets/PL_Clean.png" width="92%"></center> | -->
## Training & Inference
Detailed instructions can be found at [https://github.com/juyongjiang/CodeUp](https://github.com/juyongjiang/CodeUp).
## Citation
If you use the data or code in this repo, please cite the repo.
```
@misc{codeup,
author = {Juyong Jiang and Sunghun Kim},
title = {CodeUp: A Multilingual Code Generation Llama2 Model with Parameter-Efficient Instruction-Tuning},
year = {2023},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {\url{https://github.com/juyongjiang/CodeUp}},
}
```
Naturally, you should also cite the original LLaMA V1 [1] & V2 paper [2], and the Self-Instruct paper [3], and the LoRA paper [4], and the [Stanford Alpaca repo](https://github.com/tatsu-lab/stanford_alpaca), and [Alpaca-LoRA repo](https://github.com/tloen/alpaca-lora), and [Code Alpaca repo](https://github.com/sahil280114/codealpaca), and [PEFT](https://github.com/huggingface/peft).
|
FunPang/llama-2-7b-guanaco-instruct-sharded-finetuned-medical-conversational
|
FunPang
| 2023-10-06T06:58:55Z | 0 | 0 | null |
[
"generated_from_trainer",
"base_model:guardrail/llama-2-7b-guanaco-instruct-sharded",
"base_model:finetune:guardrail/llama-2-7b-guanaco-instruct-sharded",
"license:apache-2.0",
"region:us"
] | null | 2023-10-06T06:17:38Z |
---
license: apache-2.0
base_model: guardrail/llama-2-7b-guanaco-instruct-sharded
tags:
- generated_from_trainer
model-index:
- name: llama-2-7b-guanaco-instruct-sharded-finetuned-medical-conversational
results: []
---
<!-- 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-2-7b-guanaco-instruct-sharded-finetuned-medical-conversational
This model is a fine-tuned version of [guardrail/llama-2-7b-guanaco-instruct-sharded](https://huggingface.co/guardrail/llama-2-7b-guanaco-instruct-sharded) 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: 2
- total_train_batch_size: 2
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.03
- training_steps: 320
### Training results
### Framework versions
- Transformers 4.34.0
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.14.0
|
cy24/hr_resume_screening
|
cy24
| 2023-10-06T06:58:46Z | 0 | 0 | null |
[
"en",
"license:apache-2.0",
"region:us"
] | null | 2023-10-06T06:57:23Z |
---
license: apache-2.0
language:
- en
---
|
julep-ai/dfe-base-en
|
julep-ai
| 2023-10-06T06:50:44Z | 7 | 3 |
sentence-transformers
|
[
"sentence-transformers",
"pytorch",
"bert",
"feature-extraction",
"sentence-similarity",
"en",
"dataset:julep-ai/dfe-stacked_samsum",
"license:mit",
"autotrain_compatible",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] |
sentence-similarity
| 2023-09-22T17:36:30Z |
---
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
license: mit
datasets:
- julep-ai/dfe-stacked_samsum
language:
- en
library_name: sentence-transformers
---
# DFE (Dialog Fact Encoder)
This is a [sentence-transformers](https://www.SBERT.net) model: It maps "dialog" & "facts" to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.
The Dialog Fact Encoder (DFE) is an embedding model trained to capture semantic relevance between conversational dialog turns and factual statements. It builds upon the BGE embedding model by adding a merge layer that transforms the embeddings based on whether the input text is a "dialog" or "fact".
Specifically, dialog inputs pass through additional dense layers to project the embeddings into a space optimized for comparing dialog turns. Similarly, factual inputs pass through separate dense layers to transform the embeddings for relevance matching against dialogs.
This allows DFE to embed dialogs and facts into a joint relevance space without needing to generate explicit search queries. DFE enables low-latency approximate matching of relevant facts to dialog turns, while avoiding the high computational cost of query generation models.
The model was trained using a triplet loss to pull dialog embeddings closer to relevant fact embeddings, while pushing non-relevant pairs further apart. This helps the model learn the nuanced semantics needed to assess relevance between dialogs and facts.
DFE provides an efficient way to embed variable conversational dialog into a relevance space with factual statements. This enables real-time semantic search over knowledge without expensive query generation.
> DFE is permissively licensed under the MIT license.
## Usage (Sentence-Transformers)
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
```
pip install -U sentence-transformers
```
Then you can use the model like this:
```python
from sentence_transformers import SentenceTransformer
dialog = """
Diwank: Hey, what are we eating for dinner today?
Ishita: Already? I thought we just ate lol
Diwank: Yeah, some of us work hard and get hungy
Ishita: Okay, what do you want to eat then?
Diwank: I want to eat out but I am thinking of something light.
""".strip()
facts = [
"Diwank likes Sushi.",
"Ishita does not like unnecessarily-pricey places restaurants",
"Diwank likes cooking.",
"Ishita is terrible at cooking.",
"Diwank loves to eat Ishita's head.",
]
model = SentenceTransformer("julep-ai/dfe-base-en")
dialog_embeddings = model.encode({"dialog": dialog})
fact_embeddings = model.encode([{"fact": fact} for fact in facts])
```
## Background
So what's _Dialog-Fact Encoder_?
It is a new model trained by the Julep AI team to match (possibly) relevant facts to a dialog. It is an embedding model which means that it takes text as an input and outputs an embedding vector (a list of float numbers). In DFE's case, it takes an extra parameter "type" which can be either "dialog" or "fact".
A regular embedding model like openai's text-embedding-ada-2 does not distinguish between different types and gives vectors that can then be used for calculating similarity. These models are great for building search because comparing vectors is relatively cheap so for a database (of say product descriptions), you can compute vectors for every row beforehand and then when a query comes in (like: "Winter coats for women"), calculate the query's embeddings and find items using vector similarity.
Unfortunately, this does not work for dialog because conversational statements and turns within a dialog are typically not in the format of a "query". Take this case for example:
**Database**:
1. Diwank likes Sushi.
2. Ishita does not like unnecessarily-pricey places restaurants
3. Diwank likes cooking.
4. Ishita is terrible at cooking.
5. Diwank loves to eat Ishita's head.
**Dialog**:
> Diwank: Hey, what are we eating for dinner today?
> Ishita: Already? I thought we just ate lol
> Diwank: Yeah, some of us work hard and get hungy
> Ishita: Okay, what do you want to eat then?
> Diwank: I want to eat out but I am thinking of something light.
Now, a text/vector/hybrid search would probably match all 5 facts to this conversation but, as you can see, only facts 1 and 2 are relevant. The only way to get the correct fact, right now, is to ask an LLM like gpt-3.5 to "generate a query" for querying the database and then using that for similarity. Unfortunately, there are three big problems with that:
- It adds latency and cost.
- You have to figure out "when" to run this process and retrieve (which is hard).
- The prompt for generating the query will have to be customized for every use case because the prompt has to "know" what is "query-able". So for example, in this case above, we would have to specify that you can write a query to search preferences of Diwank and Ishita from the database.
Here's where DFE comes in. The insight here is that embeddings for a dialog have meaningful information to distinguish whether a fact is relevant or not (that is exactly how we can tell that Fact 1 and 2 are relevant and others are not in the example above because we can "see" the meaning of the dialog). Normal embedding models are only interested in "overall similarity" and they'd miss this nuance, especially for details that were NOT present in the dialog directly (for example, fact 1 mentions Sushi whereas no food items are specifically mentioned in the dialog).
So, if this information is already there in theory, how can we learn to connect embeddings of "facts" and "dialogues" based on relevance? DFE is trained to do exactly that. DFE is about learning this "relevance" transformation of a dialog so it matches similar facts.
DFE is a built upon BGE (currently the best state-of-the-art model for embeddings). It has the base embeddings from the original BGE model and added dense layers. The base BGE model is actually frozen and left completely unchanged because it already knows how to turn a passage into an embedding very well. We add the new layers to learn how to "transform" the input depending on what kind of passage it is (a dialog or a fact) in a way that "relevant" stuff is closer together in the embedding space.
This solves all of the three problems from the "query generation" method from earlier. Instead of generating a query using an LLM, you can store facts with their DFE embeddings in the database beforehand and then just embed the dialog using DFE and match. Since this operation is so much faster, you can basically do this on every turn without much hassle.
The "query generation" method is still far superior in quality but is too prohibitive (costly + slow) in normal circumstances and DFE solves that. :)
## Technical details
It inherits the base BERT model and pooling layer from BGE to generate 768-dimensional embeddings for input text.
DFE then adds an Asymmetric projection layer with separate dense layers for "dialog" and "fact" inputs:
1. Dialog inputs pass through 2x1536D tanh layers, a dropout layer, and another 1536D tanh layer before projecting back to 768 dimensions.
2. Fact inputs pass through similar 1536D tanh layers with dropout before projecting back to 768D.
3. This asymmetric architecture allows specialization of the embeddings for relevance matching between dialogs and facts.
DFE is trained with a triplet loss using the TripletDistanceMetric.EUCLIDEAN distance function and a margin of 5. It pulls dialog embeddings closer to positively matched fact embeddings, while pushing non-relevant pairs beyond the margin.
This approach teaches DFE to transform dialog and fact embeddings into a joint relevance space optimized for low-latency semantic matching. The specialized projections allow fast approximation of relevant facts for conversational dialog turns.
## Dataset
The model was trained on a custom dataset [julep-ai/dfe-stacked_samsum](https://huggingface.co/datasets/julep-ai/dfe-stacked_samsum) that we created from [stacked-summaries/stacked-samsum-1024](https://huggingface.co/datasets/stacked-summaries/stacked-samsum-1024) by:
1. Extracting summaries for corresponding dialogs to emulate "facts"
2. Then truncating the dialogs to emulate "missing information"
3. And then augmenting the dialogs using LLMs to emulate "additional information"
## Training
Training code is available in the notebook [`training.ipynb`](https://huggingface.co/julep-ai/dfe-base-en/blob/main/training.ipynb)
The model was trained with the parameters:
**Loss**:
`sentence_transformers.losses.TripletLoss.TripletLoss` with parameters:
```
{'distance_metric': 'TripletDistanceMetric.EUCLIDEAN', 'triplet_margin': 5}
```
Parameters of the fit()-Method:
```
{
"epochs": 12,
"evaluation_steps": 0,
"evaluator": "sentence_transformers.evaluation.TripletEvaluator.TripletEvaluator",
"max_grad_norm": 1,
"optimizer_class": "<class 'lion_pytorch.lion_pytorch.Lion'>",
"optimizer_params": {
"lr": 0.0001,
"weight_decay": 0.01
},
"scheduler": "WarmupCosine",
"steps_per_epoch": null,
"warmup_steps": 100,
"weight_decay": 0.01
}
```
## Evaluation Results
<!--- Describe how your model was evaluated -->
TBD
## Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 512, 'do_lower_case': True}) with Transformer model: BertModel
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
(2): Asym(
(dialog-0): Dense({'in_features': 768, 'out_features': 1536, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh'})
(dialog-1): Dense({'in_features': 1536, 'out_features': 1536, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh'})
(dialog-2): Dropout(
(dropout_layer): Dropout(p=0.1, inplace=False)
)
(dialog-3): Dense({'in_features': 1536, 'out_features': 1536, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh'})
(dialog-4): Dense({'in_features': 1536, 'out_features': 768, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh'})
(fact-0): Dense({'in_features': 768, 'out_features': 1536, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh'})
(fact-1): Dense({'in_features': 1536, 'out_features': 1536, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh'})
(fact-2): Dropout(
(dropout_layer): Dropout(p=0.1, inplace=False)
)
(fact-3): Dense({'in_features': 1536, 'out_features': 1536, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh'})
(fact-4): Dense({'in_features': 1536, 'out_features': 768, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh'})
)
)
```
## Citing & Authors
```
Diwank Singh Tomer, Julep AI Inc. Dialog Fact Encoder (DFE). https://julep.ai (2023).
```
|
wide-video/piper-voices-v1.0.0
|
wide-video
| 2023-10-06T06:48:08Z | 0 | 0 | null |
[
"onnx",
"license:mit",
"region:us"
] | null | 2023-10-06T06:48:06Z |
---
license: mit
---
Voices for [Piper](https://github.com/rhasspy/piper) text to speech system.
For checkpoints that you can use to [train your own voices](https://github.com/rhasspy/piper/blob/master/TRAINING.md), see [piper-checkpoints](https://huggingface.co/datasets/rhasspy/piper-checkpoints/tree/main)
|
Norod78/sdxl-futurama-style-lora
|
Norod78
| 2023-10-06T06:47:05Z | 29 | 2 |
diffusers
|
[
"diffusers",
"text-to-image",
"stable-diffusion",
"lora",
"style",
"cartoon",
"futurama",
"base_model:stabilityai/stable-diffusion-xl-base-1.0",
"base_model:adapter:stabilityai/stable-diffusion-xl-base-1.0",
"license:other",
"region:us"
] |
text-to-image
| 2023-10-06T06:47:03Z |
---
license: other
tags:
- text-to-image
- stable-diffusion
- lora
- diffusers
- style
- cartoon
- futurama
base_model: stabilityai/stable-diffusion-xl-base-1.0
instance_prompt: FuturamaStyle
widget:
- text: "the girl with a pearl earring, Very detailed, clean, high quality, sharp image, FuturamaStyle "
- text: "van gogh self portrait, FuturamaStyle , Very detailed, clean, high quality, sharp image"
- text: "An evil genius wearing a lab coat is working on an Apple MacBook laptop, FuturamaStyle , Very detailed, clean, high quality, sharp image"
- text: "Dora the space explorer, FuturamaStyle , Very detailed, clean, high quality, sharp image"
- text: "enchanted landscape, FuturamaStyle , Very detailed, clean, high quality, sharp image"
- text: "A full body shot of Gal Gadot as wonderwoman, FuturamaStyle , Very detailed, clean, high quality, sharp image"
- text: "Superman, FuturamaStyle , Very detailed, clean, high quality, sharp image"
- text: "Turanga Leela ass the sexy captain of the Planet Express Ship of Futurama, FuturamaStyle , Very detailed, clean, high quality, sharp image"
- text: "The head of Snoop Dogg in a jar, FuturamaStyle , Very detailed, clean, high quality, sharp image"
- text: "The planet express spaceship of Futurama, FuturamaStyle , Very detailed, clean, high quality, sharp image"
---
# SDXL Futurama-Style Lora
> the girl with a pearl earring, Very detailed, clean, high quality, sharp image, FuturamaStyle
([CivitAI](https://civitai.com/models/153437))
<p>My attempt at creating a "Futurama Style" SDXL LoRA, with limited success.</p><p>Use <strong>FuturamaStyle </strong>in your prompts.</p><p>🤖</p><p></p>
## Image examples for the model:
> van gogh self portrait, FuturamaStyle , Very detailed, clean, high quality, sharp image
> An evil genius wearing a lab coat is working on an Apple MacBook laptop, FuturamaStyle , Very detailed, clean, high quality, sharp image
> Dora the space explorer, FuturamaStyle , Very detailed, clean, high quality, sharp image
> enchanted landscape, FuturamaStyle , Very detailed, clean, high quality, sharp image
> A full body shot of Gal Gadot as wonderwoman, FuturamaStyle , Very detailed, clean, high quality, sharp image
> Superman, FuturamaStyle , Very detailed, clean, high quality, sharp image
> Turanga Leela ass the sexy captain of the Planet Express Ship of Futurama, FuturamaStyle , Very detailed, clean, high quality, sharp image
> The head of Snoop Dogg in a jar, FuturamaStyle , Very detailed, clean, high quality, sharp image
> The planet express spaceship of Futurama, FuturamaStyle , Very detailed, clean, high quality, sharp image
|
BEE-spoke-data/mobilebert-uncased-title2genre
|
BEE-spoke-data
| 2023-10-06T06:37:39Z | 113 | 1 |
transformers
|
[
"transformers",
"safetensors",
"mobilebert",
"text-classification",
"dataset tools",
"books",
"book",
"genre",
"en",
"base_model:google/mobilebert-uncased",
"base_model:finetune:google/mobilebert-uncased",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-10-05T11:36:41Z |
---
license: apache-2.0
base_model: google/mobilebert-uncased
tags:
- dataset tools
- books
- book
- genre
metrics:
- f1
widget:
- text: The Quantum Chip
example_title: Science Fiction & Fantasy
- text: One Dollar's Journey
example_title: Business & Finance
- text: Timmy The Talking Tree
example_title: idk fiction
- text: The Cursed Canvas
example_title: Arts & Design
- text: Hoops and Hegel
example_title: Philosophy & Religion
- text: Overview of Streams in North Dakota
example_title: Nature
- text: Advanced Topology
example_title: Non-fiction/Math
- text: Cooking Up Love
example_title: Food & Cooking
- text: Dr. Doolittle's Extraplanatary Commute
example_title: Science & Technology
language:
- en
---
# mobilebert-uncased-title2genre
This model is a fine-tuned version of [google/mobilebert-uncased](https://huggingface.co/google/mobilebert-uncased) for multi-label classification (18 labels).
## Model description
This classifies one or more **genre** labels in a **multi-label** setting for a given book **title**.
The 'standard' way of interpreting the predictions is that the predicted labels for a given example are **only the ones with a greater than 50% probability.**
## Details
### Labels
There are 18 labels, these are already integrated into the `config.json` and should be output by the model:
```json
"id2label": {
"0": "History & Politics",
"1": "Health & Medicine",
"2": "Mystery & Thriller",
"3": "Arts & Design",
"4": "Self-Help & Wellness",
"5": "Sports & Recreation",
"6": "Non-Fiction",
"7": "Science Fiction & Fantasy",
"8": "Countries & Geography",
"9": "Other",
"10": "Nature & Environment",
"11": "Business & Finance",
"12": "Romance",
"13": "Philosophy & Religion",
"14": "Literature & Fiction",
"15": "Science & Technology",
"16": "Children & Young Adult",
"17": "Food & Cooking"
},
```
### Eval results (validation)
It achieves the following results on the evaluation set:
- Loss: 0.2658
- F1: 0.5395
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 3e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 16
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-10
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.03
- num_epochs: 10.0
### Framework versions
- Transformers 4.35.0.dev0
- Pytorch 2.0.1+cpu
- Datasets 2.14.5
- Tokenizers 0.14.0
|
robot12343/dog
|
robot12343
| 2023-10-06T06:31:27Z | 0 | 0 | null |
[
"license:bigcode-openrail-m",
"region:us"
] | null | 2023-10-06T06:31:27Z |
---
license: bigcode-openrail-m
---
|
pekoragg/empty_motion_module_v2
|
pekoragg
| 2023-10-06T06:31:12Z | 0 | 0 | null |
[
"region:us"
] | null | 2023-09-25T20:21:55Z |
Just an empty v2 anmatediff motion module with weights all set to 0
https://huggingface.co/guoyww/animatediff/tree/main
idk what for
|
dixitakriti/lora_summarization
|
dixitakriti
| 2023-10-06T06:22:13Z | 0 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-10-06T06:22:12Z |
---
library_name: peft
---
## Training procedure
### Framework versions
- PEFT 0.5.0
|
Jzuluaga/wav2vec2-xls-r-300m-en-atc-uwb-atcc-and-atcosim
|
Jzuluaga
| 2023-10-06T06:18:32Z | 206 | 7 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"wav2vec2",
"automatic-speech-recognition",
"audio",
"en-atc",
"en",
"generated_from_trainer",
"dataset:Jzuluaga/atcosim_corpus",
"dataset:Jzuluaga/uwb_atcc",
"arxiv:2203.16822",
"arxiv:2211.04054",
"license:apache-2.0",
"model-index",
"endpoints_compatible",
"region:us"
] |
automatic-speech-recognition
| 2022-11-30T07:57:14Z |
---
license: apache-2.0
language: en
datasets:
- Jzuluaga/atcosim_corpus
- Jzuluaga/uwb_atcc
tags:
- audio
- automatic-speech-recognition
- en-atc
- en
- generated_from_trainer
metrics:
- wer
model-index:
- name: wav2vec2-xls-r-300m-en-atc-uwb-atcc-atcosim
results:
- task:
type: automatic-speech-recognition
name: Speech Recognition
dataset:
type: Jzuluaga/uwb_atcc
name: UWB-ATCC dataset (Air Traffic Control Communications)
config: test
split: test
metrics:
- type: wer
value: 24.96
name: TEST WER
verified: False
- type: wer
value: 17.9
name: TEST WER (+LM)
verified: False
- task:
type: automatic-speech-recognition
name: Speech Recognition
dataset:
type: Jzuluaga/atcosim_corpus
name: ATCOSIM corpus (Air Traffic Control Communications)
config: test
split: test
metrics:
- type: wer
value: 4.09
name: TEST WER
verified: False
- type: wer
value: 2.53
name: TEST WER (+LM)
verified: False
---
# wav2vec2-xls-r-300m-en-atc-uwb-atcc-and-atcosim
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on two corpus:
- [UWB-ATCC corpus](https://huggingface.co/datasets/Jzuluaga/uwb_atcc), and
- [ATCOSIM corpus](https://huggingface.co/datasets/Jzuluaga/atcosim_corpus).
<a href="https://colab.research.google.com/github/idiap/w2v2-air-traffic/blob/main/src/eval_xlsr_atc_model.ipynb">
<img alt="GitHub" src="https://colab.research.google.com/assets/colab-badge.svg\">
</a>
<a href="https://github.com/idiap/w2v2-air-traffic">
<img alt="GitHub" src="https://img.shields.io/badge/GitHub-Open%20source-green\">
</a>
It achieves the following results on the evaluation set (two tests sets joined together: UWB-ATCC and ATCOSIM):
- Loss: 0.5595
- Wer: 0.1687
Paper: [How Does Pre-trained Wav2Vec 2.0 Perform on Domain Shifted ASR? An Extensive Benchmark on Air Traffic Control Communications](https://arxiv.org/abs/2203.16822).
Authors: Juan Zuluaga-Gomez, Amrutha Prasad, Iuliia Nigmatulina, Saeed Sarfjoo, Petr Motlicek, Matthias Kleinert, Hartmut Helmke, Oliver Ohneiser, Qingran Zhan
Abstract: Recent work on self-supervised pre-training focus</b> on leveraging large-scale unlabeled speech data to build robust end-to-end (E2E)acoustic models (AM) that can be later fine-tuned on downstream tasks e.g., automatic speech recognition (ASR). Yet, few works investigated the impact on performance when the data properties substantially differ between the pre-training and fine-tuning phases, termed domain shift. We target this scenario by analyzing the robustness of Wav2Vec 2.0 and XLS-R models on downstream ASR for a completely unseen domain, air traffic control (ATC) communications. We benchmark these two models on several open-source and challenging ATC databases with signal-to-noise ratio between 5 and 20 dB. Relative word error rate (WER) reductions between 20% to 40% are obtained in comparison to hybrid-based ASR baselines by only fine-tuning E2E acoustic models with a smaller fraction of labeled data. We analyze WERs on the low-resource scenario and gender bias carried by one ATC dataset.
Code — GitHub repository: https://github.com/idiap/w2v2-air-traffic
## Usage
You can use our Google Colab notebook to run and evaluate our model: https://github.com/idiap/w2v2-air-traffic/blob/master/src/eval_xlsr_atc_model.ipynb
## Intended uses & limitations
This model was fine-tuned on air traffic control data. We don't expect that it keeps the same performance on some others datasets, e.g., LibriSpeech or CommonVoice.
## Training and evaluation data
See Table 1 (page 3) in our paper: [How Does Pre-trained Wav2Vec 2.0 Perform on Domain Shifted ASR? An Extensive Benchmark on Air Traffic Control Communications](https://arxiv.org/abs/2203.16822). We described there the partitions of how to use our model.
- We use the UWB-ATCC + ATCOSIM corpus to fine-tune this model. You can download the raw data here:
- https://lindat.mff.cuni.cz/repository/xmlui/handle/11858/00-097C-0000-0001-CCA1-0 and,
- https://www.spsc.tugraz.at/databases-and-tools/atcosim-air-traffic-control-simulation-speech-corpus.html
- However, do not worry, we have prepared the database in `Datasets format`:
- Here, [UWB-ATCC corpus on HuggingFace](https://huggingface.co/datasets/Jzuluaga/uwb_atcc).
- Here: [ATCOSIM CORPUS on HuggingFace](https://huggingface.co/datasets/Jzuluaga/atcosim_corpus).
- If you want to prepare a database in HuggingFace format, you can follow the data loader script in: [data_loader_atc.py](https://huggingface.co/datasets/Jzuluaga/uwb_atcc/blob/main/atc_data_loader.py).
## Writing your own inference script
If you use language model, you need to install the KenLM bindings with:
```bash
conda activate your_environment
pip install https://github.com/kpu/kenlm/archive/master.zip
```
The snippet of code:
```python
from datasets import load_dataset, load_metric, Audio
import torch
from transformers import AutoModelForCTC, Wav2Vec2Processor, Wav2Vec2ProcessorWithLM
import torchaudio.functional as F
USE_LM = False
DATASET_ID = "Jzuluaga/uwb_atcc"
MODEL_ID = "Jzuluaga/wav2vec2-xls-r-300m-en-atc-uwb-atcc-and-atcosim"
# 1. Load the dataset
# we only load the 'test' partition, however, if you want to load the 'train' partition, you can change it accordingly
uwb_atcc_corpus_test = load_dataset(DATASET_ID, "test", split="test")
# 2. Load the model
model = AutoModelForCTC.from_pretrained(MODEL_ID)
# 3. Load the processors, we offer support with LM, which should yield better resutls
if USE_LM:
processor = Wav2Vec2ProcessorWithLM.from_pretrained(MODEL_ID)
else:
processor = Wav2Vec2Processor.from_pretrained(MODEL_ID)
# 4. Format the test sample
sample = next(iter(uwb_atcc_corpus_test))
file_sampling_rate = sample['audio']['sampling_rate']
# resample if neccessary
if file_sampling_rate != 16000:
resampled_audio = F.resample(torch.tensor(sample["audio"]["array"]), file_sampling_rate, 16000).numpy()
else:
resampled_audio = torch.tensor(sample["audio"]["array"]).numpy()
input_values = processor(resampled_audio, return_tensors="pt").input_values
# 5. Run the forward pass in the model
with torch.no_grad():
logits = model(input_values).logits
# get the transcription with processor
if USE_LM:
transcription = processor.batch_decode(logits.numpy()).text
else:
pred_ids = torch.argmax(logits, dim=-1)
transcription = processor.batch_decode(pred_ids)
# print the output
print(transcription)
```
# Cite us
If you use this code for your research, please cite our paper with:
```
@article{zuluaga2022how,
title={How Does Pre-trained Wav2Vec2. 0 Perform on Domain Shifted ASR? An Extensive Benchmark on Air Traffic Control Communications},
author={Zuluaga-Gomez, Juan and Prasad, Amrutha and Nigmatulina, Iuliia and Sarfjoo, Saeed and others},
journal={IEEE Spoken Language Technology Workshop (SLT), Doha, Qatar},
year={2022}
}
```
and,
```
@article{zuluaga2022bertraffic,
title={BERTraffic: BERT-based Joint Speaker Role and Speaker Change Detection for Air Traffic Control Communications},
author={Zuluaga-Gomez, Juan and Sarfjoo, Seyyed Saeed and Prasad, Amrutha and others},
journal={IEEE Spoken Language Technology Workshop (SLT), Doha, Qatar},
year={2022}
}
```
and,
```
@article{zuluaga2022atco2,
title={ATCO2 corpus: A Large-Scale Dataset for Research on Automatic Speech Recognition and Natural Language Understanding of Air Traffic Control Communications},
author={Zuluaga-Gomez, Juan and Vesel{\`y}, Karel and Sz{\"o}ke, Igor and Motlicek, Petr and others},
journal={arXiv preprint arXiv:2211.04054},
year={2022}
}
```
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 24
- eval_batch_size: 12
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- training_steps: 10000
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:-----:|:-----:|:---------------:|:------:|
| No log | 0.63 | 500 | 3.0458 | 1.0 |
| 2.9181 | 1.27 | 1000 | 1.1503 | 0.4723 |
| 2.9181 | 1.9 | 1500 | 0.8275 | 0.3500 |
| 0.7646 | 2.53 | 2000 | 0.6990 | 0.2845 |
| 0.7646 | 3.17 | 2500 | 0.5828 | 0.2509 |
| 0.5394 | 3.8 | 3000 | 0.5363 | 0.2487 |
| 0.5394 | 4.44 | 3500 | 0.5467 | 0.2171 |
| 0.4558 | 5.07 | 4000 | 0.5290 | 0.2090 |
| 0.4558 | 5.7 | 4500 | 0.4992 | 0.2046 |
| 0.3773 | 6.34 | 5000 | 0.4934 | 0.2052 |
| 0.3773 | 6.97 | 5500 | 0.4700 | 0.1983 |
| 0.3301 | 7.6 | 6000 | 0.4938 | 0.1874 |
| 0.3301 | 8.24 | 6500 | 0.5364 | 0.1893 |
| 0.2938 | 8.87 | 7000 | 0.5170 | 0.1830 |
| 0.2938 | 9.51 | 7500 | 0.5408 | 0.1815 |
| 0.2674 | 10.14 | 8000 | 0.5581 | 0.1733 |
| 0.2674 | 10.77 | 8500 | 0.5389 | 0.1719 |
| 0.24 | 11.41 | 9000 | 0.5344 | 0.1714 |
| 0.24 | 12.04 | 9500 | 0.5503 | 0.1686 |
| 0.211 | 12.67 | 10000 | 0.5595 | 0.1687 |
### Framework versions
- Transformers 4.24.0
- Pytorch 1.13.0+cu117
- Datasets 2.6.1
- Tokenizers 0.13.2
|
abeiler/NumAndAlphaInstruct
|
abeiler
| 2023-10-06T06:02:54Z | 5 | 0 |
transformers
|
[
"transformers",
"pytorch",
"llama",
"text-generation",
"generated_from_trainer",
"base_model:meta-llama/Llama-2-7b-hf",
"base_model:finetune:meta-llama/Llama-2-7b-hf",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text-generation
| 2023-10-06T01:53:03Z |
---
base_model: meta-llama/Llama-2-7b-hf
tags:
- generated_from_trainer
model-index:
- name: goatNumAndAlphaInstruct-QLORA
results: []
---
<!-- 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. -->
# goatNumAndAlphaInstruct-QLORA
This model is a fine-tuned version of [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-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: 0.0001
- 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: linear
- num_epochs: 1
### Training results
### Framework versions
- Transformers 4.33.3
- Pytorch 2.0.0
- Datasets 2.12.0
- Tokenizers 0.13.3
|
CHAEBBANG/flan-t5-large-financial-phrasebank-lora
|
CHAEBBANG
| 2023-10-06T05:52:01Z | 2 | 0 |
peft
|
[
"peft",
"arxiv:1910.09700",
"base_model:google/flan-t5-large",
"base_model:adapter:google/flan-t5-large",
"region:us"
] | null | 2023-10-06T05:36:22Z |
---
library_name: peft
base_model: google/flan-t5-large
---
# 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]
- **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 Data 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 Data 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]
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: True
- load_in_4bit: False
- 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: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
### Framework versions
- PEFT 0.6.0.dev0
|
shengqin/bloomz-560m_PROMPT_TUNING_CAUSAL_LM
|
shengqin
| 2023-10-06T05:46:46Z | 0 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-09-06T03:18:41Z |
---
library_name: peft
---
## Training procedure
### Framework versions
- PEFT 0.5.0
|
ArmelRandy/llama-2-13b-precious-380
|
ArmelRandy
| 2023-10-06T05:46:08Z | 0 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-10-06T05:45:03Z |
---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: True
- load_in_4bit: False
- 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: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
### Framework versions
- PEFT 0.5.0
|
soBeauty/20231005-3-bert-base-multilingual-cased-new
|
soBeauty
| 2023-10-06T05:36:21Z | 194 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"fill-mask",
"generated_from_trainer",
"base_model:google-bert/bert-base-multilingual-cased",
"base_model:finetune:google-bert/bert-base-multilingual-cased",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2023-10-06T00:35:15Z |
---
license: apache-2.0
base_model: bert-base-multilingual-cased
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: 20231005-3-bert-base-multilingual-cased-new
results: []
---
<!-- 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. -->
# 20231005-3-bert-base-multilingual-cased-new
This model is a fine-tuned version of [bert-base-multilingual-cased](https://huggingface.co/bert-base-multilingual-cased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Accuracy: 0.6077
- Loss: 1.7371
## 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: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 20
### Training results
| Training Loss | Epoch | Step | Accuracy | Validation Loss |
|:-------------:|:-----:|:----:|:--------:|:---------------:|
| 2.9237 | 1.82 | 200 | 0.4526 | 2.4628 |
| 2.3796 | 3.64 | 400 | 0.4668 | 2.4166 |
| 2.2074 | 5.45 | 600 | 0.5011 | 2.1639 |
| 2.1154 | 7.27 | 800 | 0.5333 | 1.9224 |
| 1.9577 | 9.09 | 1000 | 0.5607 | 1.8915 |
| 1.8846 | 10.91 | 1200 | 0.6009 | 1.5682 |
| 1.8251 | 12.73 | 1400 | 0.6024 | 1.6171 |
| 1.7344 | 14.55 | 1600 | 0.5923 | 1.6795 |
| 1.7236 | 16.36 | 1800 | 0.6191 | 1.5585 |
| 1.7099 | 18.18 | 2000 | 0.6241 | 1.6113 |
| 1.6309 | 20.0 | 2200 | 0.6077 | 1.7371 |
### Framework versions
- Transformers 4.34.0
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.14.0
|
madatnlp/km-bert
|
madatnlp
| 2023-10-06T05:22:27Z | 1,567 | 5 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"bert",
"feature-extraction",
"biology",
"ko",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] |
feature-extraction
| 2023-07-04T03:04:39Z |
---
language:
- ko
tags:
- biology
---
# Model Card for Model ID
이 모델은 KM-BERT의 모델을 다운 받아, 임의로 허깅페이스 모델의 형태로 사용하기 좋게 변환한 모델 입니다.
<br>
모델은 그대로이고, 토크나이저는 논문에 나와 있는대로, snunlp/KR-BERT-char16424 것을 가져왔습니다
<br>
KM-BERT를 사용해볼 수 있게 공유해주신 저자분께 감사드립니다.
<br>
문제 발생 시 수정하거나, 또는 저자님의 요청시 허브에서 내릴 수도 있습니다.
<br>
## Model Details
bing 활용 요약:
이 문서는 다음과 같은 내용을 다룹니다:
한국어 의료 자연어 처리를 위한 사전 훈련된 BERT 모델. <br>
이 모델은 의료 분야에서 사용되는 한국어 텍스트를 분석하기 위해 KR-BERT 모델을 기반으로 한국어 의료 코퍼스로 사전 훈련한 KM-BERT라고 합니다. <br>
이 모델은 의료 교과서, 건강 정보 뉴스, 의료 연구 논문 등 세 가지 유형의 문서로 구성된 약 1억 1천 6백만 단어의 한국어 의료 코퍼스를 사용했습니다. <br>
내부 평가와 외부 평가. <br>
이 모델은 MLM과 NSP라는 두 가지 비지도 사전 훈련 과제를 수행하고, 새로운 한국어 의료 코퍼스와 MedSTS 데이터셋을 이용하여 언어 이해 능력을 내부 평가했습니다. <br>
또한, 한국어 의료 의미론적 텍스트 유사도(MedSTS) 데이터셋과 한국어 의료 개체 인식(NER) 데이터셋을 이용하여 하류 과제에 대한 성능을 외부 평가했습니다. <br>
결과적으로, KM-BERT는 기존의 언어 모델들보다 우수한 성능을 보였습니다. <br>
언어별 및 분야별 관점. <br>
M-BERT는 다양한 언어에 대한 NLP 과제에 높은 전이성을 보이지만, 언어별 BERT 모델들이 M-BERT보다 더 나은 성능을 보인다는 것이 증명되었습니다. <br>
1. 한국어는 영어와 같은 인도유럽어와는 다른 단어 순서를 가지고 있기 때문에, 한국어에 특화된 BERT 모델이 필요합니다. <br>
2. 또한, SciBERT나 LEGAL-BERT와 같은 분야별 BERT 임베딩이 BERT보다 전문 용어와 용법에 대한 이해력을 향상시킨다는 것이 밝혀졌습니다. <br>
3. 이러한 연구 결과들은 한국어 의료 분야에 적합한 BERT 모델 개발의 필요성과 효과성을 보여줍니다. <br>
### Model Sources [optional]
- **Repository:** https://github.com/KU-RIAS/KM-BERT-Korean-Medical-BERT <br>
- **Paper:** https://www.nature.com/articles/s41598-022-17806-8 <br>
## Citation [optional]
@article{KMBERT, <br>
title={KM-BERT: A Pre-trained BERT for Korean Medical Natural Language Processing}, <br>
author={TBD}, <br>
year={TBD}, <br>
journal={TBD}, <br>
volume={TBD} <br>
}
|
LarryAIDraw/rosaria_genshin
|
LarryAIDraw
| 2023-10-06T05:17:21Z | 0 | 0 | null |
[
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-08-25T03:53:43Z |
---
license: creativeml-openrail-m
---
https://civitai.com/models/130977/rosaria-genshin-impact
|
euiik/test_segformer
|
euiik
| 2023-10-06T04:36:02Z | 31 | 0 |
transformers
|
[
"transformers",
"tf",
"segformer",
"generated_from_keras_callback",
"base_model:nvidia/mit-b0",
"base_model:finetune:nvidia/mit-b0",
"license:other",
"endpoints_compatible",
"region:us"
] | null | 2023-10-06T04:33:34Z |
---
license: other
base_model: nvidia/mit-b0
tags:
- generated_from_keras_callback
model-index:
- name: test_segformer
results: []
---
<!-- 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. -->
# test_segformer
This model is a fine-tuned version of [nvidia/mit-b0](https://huggingface.co/nvidia/mit-b0) on an unknown dataset.
It achieves the following results on the evaluation set:
- Train Loss: 0.8965
- Validation Loss: 1.1488
- 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': 'Adam', 'weight_decay': None, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': True, 'is_legacy_optimizer': False, 'learning_rate': 6e-05, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False}
- training_precision: float32
### Training results
| Train Loss | Validation Loss | Epoch |
|:----------:|:---------------:|:-----:|
| 0.8965 | 1.1488 | 0 |
### Framework versions
- Transformers 4.33.2
- TensorFlow 2.13.0
- Tokenizers 0.13.3
|
Brazilia/Pelicikudbhd
|
Brazilia
| 2023-10-06T04:18:37Z | 0 | 0 | null |
[
"license:creativeml-openrail-m",
"region:us"
] | null | 2023-10-06T04:15:06Z |
---
license: creativeml-openrail-m
---
|
Aradwy/Shastoon_aradwy_250Epoch
|
Aradwy
| 2023-10-06T04:17:59Z | 0 | 0 | null |
[
"разговорный",
"Shastoon",
"TikTok:@_aradwy",
"aicover",
"twitter:@_aradwy",
"Аи кавер",
"Антон Шастун",
"improvisation",
"импровизаторы",
"импровизация",
"стс",
"реп",
"Aradwy",
"ru",
"license:unknown",
"region:us"
] | null | 2023-09-21T14:13:30Z |
---
license: unknown
language:
- ru
tags:
- разговорный
- Shastoon
- TikTok:@_aradwy
- aicover
- twitter:@_aradwy
- Аи кавер
- Антон Шастун
- improvisation
- импровизаторы
- импровизация
- стс
- реп
- Aradwy
---
|
aierwiki/chatglm2-baby
|
aierwiki
| 2023-10-06T04:16:23Z | 134 | 0 |
transformers
|
[
"transformers",
"pytorch",
"chatglm",
"custom_code",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null | 2023-10-05T10:09:48Z |
---
license: apache-2.0
---
chatglm2-baby是为了方便在CPU机器上调试代码，将chatglm2-6b的模型配置参数大幅度降低，代码和官方保持一致。当前可以在cpu上快速跑通推理流程。
|
kwwww/test_64
|
kwwww
| 2023-10-06T03:53:52Z | 0 | 0 | null |
[
"pytorch",
"generated_from_trainer",
"dataset:bionlp2004",
"license:apache-2.0",
"region:us"
] | null | 2023-10-05T03:24:01Z |
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- bionlp2004
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: test_64
results: []
---
<!-- 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. -->
# test_64
This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the bionlp2004 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5893
- Precision: 0.8007
- Recall: 0.8210
- F1: 0.8107
- Accuracy: 0.9467
## 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: 0.001
- 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: 1000
### Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|:-------------:|:------:|:-------:|:---------------:|:---------:|:------:|:------:|:--------:|
| 0.2129 | 1.0 | 1039 | 0.1893 | 0.7163 | 0.7463 | 0.7310 | 0.9356 |
| 0.1846 | 2.0 | 2078 | 0.1721 | 0.7311 | 0.8084 | 0.7678 | 0.9417 |
| 0.1704 | 3.0 | 3117 | 0.1702 | 0.7403 | 0.7920 | 0.7653 | 0.9441 |
| 0.1576 | 4.0 | 4156 | 0.1687 | 0.7521 | 0.8190 | 0.7841 | 0.9448 |
| 0.1505 | 5.0 | 5195 | 0.1719 | 0.7438 | 0.7952 | 0.7687 | 0.9449 |
| 0.1446 | 6.0 | 6234 | 0.1770 | 0.7403 | 0.8080 | 0.7727 | 0.9441 |
| 0.1341 | 7.0 | 7273 | 0.1784 | 0.7472 | 0.8363 | 0.7893 | 0.9460 |
| 0.1293 | 8.0 | 8312 | 0.1764 | 0.7657 | 0.8165 | 0.7903 | 0.9454 |
| 0.1303 | 9.0 | 9351 | 0.1678 | 0.7586 | 0.8082 | 0.7826 | 0.9462 |
| 0.1219 | 10.0 | 10390 | 0.1700 | 0.7686 | 0.8367 | 0.8012 | 0.9477 |
| 0.1165 | 11.0 | 11429 | 0.1845 | 0.7471 | 0.8221 | 0.7828 | 0.9456 |
| 0.11 | 12.0 | 12468 | 0.1794 | 0.7614 | 0.8277 | 0.7932 | 0.9447 |
| 0.1152 | 13.0 | 13507 | 0.1855 | 0.7410 | 0.8124 | 0.7750 | 0.9442 |
| 0.1096 | 14.0 | 14546 | 0.1756 | 0.7679 | 0.8037 | 0.7854 | 0.9452 |
| 0.1073 | 15.0 | 15585 | 0.1810 | 0.7537 | 0.8129 | 0.7822 | 0.9444 |
| 0.104 | 16.0 | 16624 | 0.1822 | 0.7548 | 0.8145 | 0.7835 | 0.9447 |
| 0.0996 | 17.0 | 17663 | 0.1871 | 0.7800 | 0.8109 | 0.7952 | 0.9462 |
| 0.0976 | 18.0 | 18702 | 0.2029 | 0.7601 | 0.8057 | 0.7822 | 0.9436 |
| 0.0929 | 19.0 | 19741 | 0.1986 | 0.7798 | 0.7859 | 0.7829 | 0.9444 |
| 0.091 | 20.0 | 20780 | 0.1909 | 0.7744 | 0.8080 | 0.7909 | 0.9455 |
| 0.0896 | 21.0 | 21819 | 0.2079 | 0.7651 | 0.8113 | 0.7875 | 0.9450 |
| 0.0867 | 22.0 | 22858 | 0.1948 | 0.7773 | 0.8084 | 0.7925 | 0.9453 |
| 0.0836 | 23.0 | 23897 | 0.1959 | 0.7613 | 0.8237 | 0.7913 | 0.9444 |
| 0.0816 | 24.0 | 24936 | 0.2181 | 0.7655 | 0.8120 | 0.7881 | 0.9441 |
| 0.0766 | 25.0 | 25975 | 0.2009 | 0.7706 | 0.8120 | 0.7908 | 0.9445 |
| 0.0859 | 26.0 | 27014 | 0.2032 | 0.7900 | 0.8050 | 0.7974 | 0.9461 |
| 0.0818 | 27.0 | 28053 | 0.2090 | 0.7670 | 0.8152 | 0.7904 | 0.9451 |
| 0.078 | 28.0 | 29092 | 0.2251 | 0.7683 | 0.7916 | 0.7798 | 0.9429 |
| 0.0764 | 29.0 | 30131 | 0.2134 | 0.7782 | 0.8010 | 0.7894 | 0.9455 |
| 0.0728 | 30.0 | 31170 | 0.2158 | 0.7699 | 0.8394 | 0.8031 | 0.9472 |
| 0.0713 | 31.0 | 32209 | 0.2129 | 0.7690 | 0.8136 | 0.7907 | 0.9446 |
| 0.069 | 32.0 | 33248 | 0.2261 | 0.7624 | 0.8024 | 0.7819 | 0.9436 |
| 0.0697 | 33.0 | 34287 | 0.2293 | 0.7648 | 0.8190 | 0.7910 | 0.9436 |
| 0.0673 | 34.0 | 35326 | 0.2190 | 0.7731 | 0.8046 | 0.7886 | 0.9447 |
| 0.0643 | 35.0 | 36365 | 0.2388 | 0.7654 | 0.8077 | 0.7859 | 0.9431 |
| 0.0602 | 36.0 | 37404 | 0.2319 | 0.7525 | 0.8115 | 0.7809 | 0.9440 |
| 0.063 | 37.0 | 38443 | 0.2226 | 0.7824 | 0.8127 | 0.7973 | 0.9448 |
| 0.0582 | 38.0 | 39482 | 0.2417 | 0.7813 | 0.7965 | 0.7888 | 0.9447 |
| 0.0657 | 39.0 | 40521 | 0.2430 | 0.7847 | 0.7927 | 0.7887 | 0.9431 |
| 0.0634 | 40.0 | 41560 | 0.2464 | 0.7728 | 0.8116 | 0.7917 | 0.9446 |
| 0.0609 | 41.0 | 42599 | 0.2421 | 0.7758 | 0.8151 | 0.7949 | 0.9451 |
| 0.0588 | 42.0 | 43638 | 0.2449 | 0.7657 | 0.8167 | 0.7903 | 0.9436 |
| 0.0578 | 43.0 | 44677 | 0.2339 | 0.7724 | 0.8111 | 0.7913 | 0.9448 |
| 0.0578 | 44.0 | 45716 | 0.2329 | 0.7718 | 0.8230 | 0.7966 | 0.9460 |
| 0.0551 | 45.0 | 46755 | 0.2458 | 0.7706 | 0.8001 | 0.7851 | 0.9431 |
| 0.0552 | 46.0 | 47794 | 0.2573 | 0.7821 | 0.7983 | 0.7901 | 0.9446 |
| 0.0532 | 47.0 | 48833 | 0.2313 | 0.7775 | 0.8062 | 0.7916 | 0.9460 |
| 0.052 | 48.0 | 49872 | 0.2848 | 0.7666 | 0.8075 | 0.7865 | 0.9420 |
| 0.052 | 49.0 | 50911 | 0.2535 | 0.7688 | 0.8095 | 0.7886 | 0.9447 |
| 0.0493 | 50.0 | 51950 | 0.2558 | 0.7765 | 0.8152 | 0.7954 | 0.9439 |
| 0.0495 | 51.0 | 52989 | 0.2525 | 0.7756 | 0.8104 | 0.7926 | 0.9442 |
| 0.054 | 52.0 | 54028 | 0.2570 | 0.7821 | 0.8118 | 0.7967 | 0.9447 |
| 0.0522 | 53.0 | 55067 | 0.2601 | 0.7932 | 0.8046 | 0.7989 | 0.9452 |
| 0.0517 | 54.0 | 56106 | 0.2734 | 0.7660 | 0.7924 | 0.7790 | 0.9414 |
| 0.0509 | 55.0 | 57145 | 0.2661 | 0.7687 | 0.7943 | 0.7813 | 0.9424 |
| 0.0511 | 56.0 | 58184 | 0.2528 | 0.7696 | 0.8010 | 0.7850 | 0.9427 |
| 0.049 | 57.0 | 59223 | 0.2567 | 0.7837 | 0.7961 | 0.7899 | 0.9442 |
| 0.0479 | 58.0 | 60262 | 0.2603 | 0.7842 | 0.7898 | 0.7870 | 0.9410 |
| 0.0461 | 59.0 | 61301 | 0.2435 | 0.7757 | 0.7933 | 0.7844 | 0.9423 |
| 0.044 | 60.0 | 62340 | 0.2474 | 0.7815 | 0.8037 | 0.7924 | 0.9444 |
| 0.0446 | 61.0 | 63379 | 0.2580 | 0.7713 | 0.8019 | 0.7863 | 0.9448 |
| 0.0456 | 62.0 | 64418 | 0.2757 | 0.7803 | 0.8131 | 0.7964 | 0.9446 |
| 0.0425 | 63.0 | 65457 | 0.2570 | 0.7676 | 0.8005 | 0.7837 | 0.9432 |
| 0.0413 | 64.0 | 66496 | 0.2668 | 0.7876 | 0.8098 | 0.7985 | 0.9452 |
| 0.047 | 65.0 | 67535 | 0.2730 | 0.7784 | 0.8179 | 0.7977 | 0.9450 |
| 0.0463 | 66.0 | 68574 | 0.2648 | 0.7914 | 0.8091 | 0.8002 | 0.9449 |
| 0.0446 | 67.0 | 69613 | 0.2635 | 0.7846 | 0.7961 | 0.7903 | 0.9438 |
| 0.0443 | 68.0 | 70652 | 0.2662 | 0.7739 | 0.8122 | 0.7925 | 0.9450 |
| 0.0434 | 69.0 | 71691 | 0.2804 | 0.7767 | 0.7916 | 0.7841 | 0.9411 |
| 0.0436 | 70.0 | 72730 | 0.2910 | 0.7705 | 0.8158 | 0.7925 | 0.9453 |
| 0.0417 | 71.0 | 73769 | 0.2640 | 0.7774 | 0.8012 | 0.7891 | 0.9426 |
| 0.0408 | 72.0 | 74808 | 0.2705 | 0.7802 | 0.8091 | 0.7944 | 0.9440 |
| 0.0417 | 73.0 | 75847 | 0.2744 | 0.7812 | 0.8203 | 0.8002 | 0.9460 |
| 0.0393 | 74.0 | 76886 | 0.2844 | 0.7719 | 0.7994 | 0.7854 | 0.9421 |
| 0.0389 | 75.0 | 77925 | 0.2861 | 0.7611 | 0.8187 | 0.7888 | 0.9429 |
| 0.0385 | 76.0 | 78964 | 0.2745 | 0.7848 | 0.7992 | 0.7919 | 0.9436 |
| 0.0422 | 77.0 | 80003 | 0.2776 | 0.7732 | 0.8111 | 0.7917 | 0.9438 |
| 0.0387 | 78.0 | 81042 | 0.2895 | 0.7819 | 0.7920 | 0.7869 | 0.9436 |
| 0.0384 | 79.0 | 82081 | 0.2769 | 0.7740 | 0.8059 | 0.7896 | 0.9442 |
| 0.0393 | 80.0 | 83120 | 0.2880 | 0.7739 | 0.8142 | 0.7935 | 0.9444 |
| 0.0365 | 81.0 | 84159 | 0.2819 | 0.7759 | 0.7931 | 0.7844 | 0.9425 |
| 0.0386 | 82.0 | 85198 | 0.2857 | 0.7784 | 0.8014 | 0.7897 | 0.9434 |
| 0.0375 | 83.0 | 86237 | 0.2838 | 0.7776 | 0.8086 | 0.7928 | 0.9440 |
| 0.0368 | 84.0 | 87276 | 0.2726 | 0.7781 | 0.8006 | 0.7892 | 0.9434 |
| 0.0367 | 85.0 | 88315 | 0.2799 | 0.7820 | 0.8035 | 0.7926 | 0.9434 |
| 0.035 | 86.0 | 89354 | 0.2941 | 0.7728 | 0.8091 | 0.7905 | 0.9441 |
| 0.0356 | 87.0 | 90393 | 0.2917 | 0.7868 | 0.8042 | 0.7954 | 0.9444 |
| 0.0336 | 88.0 | 91432 | 0.2893 | 0.7813 | 0.8185 | 0.7995 | 0.9459 |
| 0.0331 | 89.0 | 92471 | 0.3002 | 0.7744 | 0.7769 | 0.7756 | 0.9387 |
| 0.0405 | 90.0 | 93510 | 0.2852 | 0.7754 | 0.8028 | 0.7889 | 0.9434 |
| 0.0368 | 91.0 | 94549 | 0.2954 | 0.7814 | 0.8066 | 0.7938 | 0.9439 |
| 0.0359 | 92.0 | 95588 | 0.2982 | 0.7810 | 0.8163 | 0.7983 | 0.9434 |
| 0.0344 | 93.0 | 96627 | 0.3042 | 0.7799 | 0.8073 | 0.7934 | 0.9442 |
| 0.0339 | 94.0 | 97666 | 0.2940 | 0.7760 | 0.8165 | 0.7957 | 0.9434 |
| 0.0355 | 95.0 | 98705 | 0.2971 | 0.7666 | 0.8048 | 0.7852 | 0.9423 |
| 0.0349 | 96.0 | 99744 | 0.2742 | 0.7702 | 0.7974 | 0.7836 | 0.9413 |
| 0.0323 | 97.0 | 100783 | 0.2890 | 0.7637 | 0.8026 | 0.7827 | 0.9426 |
| 0.032 | 98.0 | 101822 | 0.3021 | 0.7662 | 0.8071 | 0.7861 | 0.9420 |
| 0.0308 | 99.0 | 102861 | 0.2876 | 0.7851 | 0.8046 | 0.7947 | 0.9444 |
| 0.0334 | 100.0 | 103900 | 0.2822 | 0.7878 | 0.8116 | 0.7995 | 0.9455 |
| 0.0315 | 101.0 | 104939 | 0.3095 | 0.7743 | 0.8019 | 0.7879 | 0.9418 |
| 0.0327 | 102.0 | 105978 | 0.2988 | 0.7736 | 0.8109 | 0.7918 | 0.9438 |
| 0.0348 | 103.0 | 107017 | 0.3122 | 0.7862 | 0.7861 | 0.7861 | 0.9416 |
| 0.0344 | 104.0 | 108056 | 0.2892 | 0.7762 | 0.8120 | 0.7937 | 0.9443 |
| 0.0327 | 105.0 | 109095 | 0.3022 | 0.7707 | 0.7965 | 0.7834 | 0.9424 |
| 0.034 | 106.0 | 110134 | 0.2841 | 0.7814 | 0.7936 | 0.7875 | 0.9428 |
| 0.0321 | 107.0 | 111173 | 0.2922 | 0.7726 | 0.8120 | 0.7918 | 0.9430 |
| 0.0323 | 108.0 | 112212 | 0.3233 | 0.7747 | 0.8057 | 0.7899 | 0.9432 |
| 0.0304 | 109.0 | 113251 | 0.3021 | 0.7716 | 0.7970 | 0.7841 | 0.9411 |
| 0.0305 | 110.0 | 114290 | 0.2870 | 0.7862 | 0.8097 | 0.7977 | 0.9457 |
| 0.0297 | 111.0 | 115329 | 0.3124 | 0.7817 | 0.8093 | 0.7953 | 0.9436 |
| 0.0328 | 112.0 | 116368 | 0.3088 | 0.7802 | 0.7952 | 0.7877 | 0.9422 |
| 0.0298 | 113.0 | 117407 | 0.2769 | 0.7857 | 0.8161 | 0.8006 | 0.9458 |
| 0.0284 | 114.0 | 118446 | 0.3048 | 0.7690 | 0.8154 | 0.7915 | 0.9435 |
| 0.0274 | 115.0 | 119485 | 0.3100 | 0.7927 | 0.7918 | 0.7923 | 0.9436 |
| 0.0304 | 116.0 | 120524 | 0.2863 | 0.7763 | 0.8010 | 0.7884 | 0.9441 |
| 0.0317 | 117.0 | 121563 | 0.3098 | 0.7751 | 0.8107 | 0.7925 | 0.9437 |
| 0.0314 | 118.0 | 122602 | 0.2933 | 0.7869 | 0.8035 | 0.7952 | 0.9442 |
| 0.0326 | 119.0 | 123641 | 0.3023 | 0.7800 | 0.7819 | 0.7809 | 0.9409 |
| 0.0277 | 120.0 | 124680 | 0.3110 | 0.7562 | 0.8012 | 0.7781 | 0.9411 |
| 0.0297 | 121.0 | 125719 | 0.2906 | 0.7857 | 0.8120 | 0.7986 | 0.9441 |
| 0.029 | 122.0 | 126758 | 0.3102 | 0.7936 | 0.8093 | 0.8014 | 0.9451 |
| 0.0289 | 123.0 | 127797 | 0.3026 | 0.7782 | 0.8228 | 0.7999 | 0.9451 |
| 0.0282 | 124.0 | 128836 | 0.2984 | 0.7740 | 0.8062 | 0.7898 | 0.9440 |
| 0.0295 | 125.0 | 129875 | 0.2900 | 0.7667 | 0.8250 | 0.7948 | 0.9461 |
| 0.0286 | 126.0 | 130914 | 0.3040 | 0.7932 | 0.7787 | 0.7859 | 0.9425 |
| 0.0274 | 127.0 | 131953 | 0.2921 | 0.7834 | 0.8039 | 0.7935 | 0.9450 |
| 0.0283 | 128.0 | 132992 | 0.3257 | 0.7713 | 0.8242 | 0.7969 | 0.9438 |
| 0.0279 | 129.0 | 134031 | 0.3063 | 0.7833 | 0.8111 | 0.7970 | 0.9442 |
| 0.0307 | 130.0 | 135070 | 0.3045 | 0.7844 | 0.8104 | 0.7972 | 0.9461 |
| 0.0293 | 131.0 | 136109 | 0.3011 | 0.7742 | 0.8021 | 0.7879 | 0.9429 |
| 0.0282 | 132.0 | 137148 | 0.3255 | 0.7764 | 0.7958 | 0.7859 | 0.9435 |
| 0.0282 | 133.0 | 138187 | 0.2971 | 0.7682 | 0.8217 | 0.7940 | 0.9428 |
| 0.027 | 134.0 | 139226 | 0.3143 | 0.7681 | 0.8111 | 0.7890 | 0.9434 |
| 0.0252 | 135.0 | 140265 | 0.3054 | 0.7645 | 0.8161 | 0.7895 | 0.9440 |
| 0.0282 | 136.0 | 141304 | 0.3010 | 0.7775 | 0.8066 | 0.7918 | 0.9441 |
| 0.0264 | 137.0 | 142343 | 0.3122 | 0.7899 | 0.8152 | 0.8024 | 0.9466 |
| 0.028 | 138.0 | 143382 | 0.2836 | 0.7811 | 0.8071 | 0.7939 | 0.9443 |
| 0.0242 | 139.0 | 144421 | 0.3166 | 0.7913 | 0.7906 | 0.7909 | 0.9430 |
| 0.025 | 140.0 | 145460 | 0.3153 | 0.7703 | 0.8194 | 0.7941 | 0.9446 |
| 0.0246 | 141.0 | 146499 | 0.2719 | 0.7853 | 0.8008 | 0.7930 | 0.9440 |
| 0.028 | 142.0 | 147538 | 0.3065 | 0.7784 | 0.8048 | 0.7914 | 0.9433 |
| 0.0272 | 143.0 | 148577 | 0.3227 | 0.7659 | 0.8005 | 0.7828 | 0.9410 |
| 0.0275 | 144.0 | 149616 | 0.3137 | 0.7726 | 0.8145 | 0.7930 | 0.9450 |
| 0.0262 | 145.0 | 150655 | 0.3089 | 0.7910 | 0.8082 | 0.7995 | 0.9453 |
| 0.0247 | 146.0 | 151694 | 0.3032 | 0.7748 | 0.8109 | 0.7924 | 0.9448 |
| 0.0253 | 147.0 | 152733 | 0.2783 | 0.7854 | 0.8115 | 0.7982 | 0.9455 |
| 0.0249 | 148.0 | 153772 | 0.3084 | 0.7852 | 0.8024 | 0.7937 | 0.9456 |
| 0.0272 | 149.0 | 154811 | 0.2993 | 0.7860 | 0.8082 | 0.7969 | 0.9456 |
| 0.0238 | 150.0 | 155850 | 0.3326 | 0.7780 | 0.8109 | 0.7941 | 0.9451 |
| 0.0256 | 151.0 | 156889 | 0.3337 | 0.7788 | 0.8052 | 0.7917 | 0.9430 |
| 0.023 | 152.0 | 157928 | 0.3001 | 0.7810 | 0.8124 | 0.7964 | 0.9453 |
| 0.0245 | 153.0 | 158967 | 0.3078 | 0.7818 | 0.7819 | 0.7818 | 0.9410 |
| 0.0254 | 154.0 | 160006 | 0.3284 | 0.7851 | 0.8163 | 0.8004 | 0.9443 |
| 0.0282 | 155.0 | 161045 | 0.3198 | 0.7827 | 0.7979 | 0.7903 | 0.9432 |
| 0.0265 | 156.0 | 162084 | 0.3110 | 0.7689 | 0.8120 | 0.7899 | 0.9440 |
| 0.025 | 157.0 | 163123 | 0.2954 | 0.7766 | 0.8151 | 0.7954 | 0.9456 |
| 0.0252 | 158.0 | 164162 | 0.3126 | 0.7815 | 0.8115 | 0.7962 | 0.9444 |
| 0.0271 | 159.0 | 165201 | 0.2926 | 0.7825 | 0.7999 | 0.7911 | 0.9431 |
| 0.0268 | 160.0 | 166240 | 0.3223 | 0.7884 | 0.8084 | 0.7983 | 0.9454 |
| 0.0236 | 161.0 | 167279 | 0.3068 | 0.7804 | 0.8088 | 0.7943 | 0.9444 |
| 0.025 | 162.0 | 168318 | 0.3086 | 0.7757 | 0.8221 | 0.7982 | 0.9440 |
| 0.0241 | 163.0 | 169357 | 0.3266 | 0.7702 | 0.8095 | 0.7894 | 0.9436 |
| 0.0213 | 164.0 | 170396 | 0.3026 | 0.7800 | 0.8079 | 0.7937 | 0.9433 |
| 0.024 | 165.0 | 171435 | 0.3311 | 0.7853 | 0.8149 | 0.7998 | 0.9448 |
| 0.0237 | 166.0 | 172474 | 0.3132 | 0.7893 | 0.8151 | 0.8020 | 0.9459 |
| 0.0246 | 167.0 | 173513 | 0.3349 | 0.7703 | 0.8152 | 0.7921 | 0.9425 |
| 0.025 | 168.0 | 174552 | 0.3099 | 0.7841 | 0.7981 | 0.7911 | 0.9430 |
| 0.0245 | 169.0 | 175591 | 0.3301 | 0.7798 | 0.7992 | 0.7894 | 0.9424 |
| 0.0218 | 170.0 | 176630 | 0.3341 | 0.7757 | 0.8122 | 0.7935 | 0.9431 |
| 0.0246 | 171.0 | 177669 | 0.3204 | 0.7870 | 0.8106 | 0.7986 | 0.9455 |
| 0.0244 | 172.0 | 178708 | 0.3327 | 0.7837 | 0.8149 | 0.7990 | 0.9438 |
| 0.0228 | 173.0 | 179747 | 0.3248 | 0.7716 | 0.8061 | 0.7884 | 0.9435 |
| 0.0219 | 174.0 | 180786 | 0.3262 | 0.7748 | 0.8170 | 0.7953 | 0.9433 |
| 0.023 | 175.0 | 181825 | 0.3176 | 0.7902 | 0.8107 | 0.8004 | 0.9448 |
| 0.0226 | 176.0 | 182864 | 0.3194 | 0.7838 | 0.8044 | 0.7940 | 0.9454 |
| 0.0216 | 177.0 | 183903 | 0.3479 | 0.7763 | 0.8138 | 0.7946 | 0.9433 |
| 0.0219 | 178.0 | 184942 | 0.3492 | 0.7770 | 0.8080 | 0.7922 | 0.9429 |
| 0.0228 | 179.0 | 185981 | 0.3373 | 0.7755 | 0.8192 | 0.7967 | 0.9446 |
| 0.0251 | 180.0 | 187020 | 0.3069 | 0.7731 | 0.8120 | 0.7921 | 0.9424 |
| 0.0237 | 181.0 | 188059 | 0.3228 | 0.7859 | 0.8057 | 0.7957 | 0.9444 |
| 0.0225 | 182.0 | 189098 | 0.3479 | 0.7684 | 0.8068 | 0.7871 | 0.9426 |
| 0.0231 | 183.0 | 190137 | 0.3362 | 0.7781 | 0.8077 | 0.7926 | 0.9449 |
| 0.0237 | 184.0 | 191176 | 0.3224 | 0.7774 | 0.8151 | 0.7958 | 0.9437 |
| 0.0193 | 185.0 | 192215 | 0.3221 | 0.7871 | 0.8084 | 0.7976 | 0.9449 |
| 0.0221 | 186.0 | 193254 | 0.3174 | 0.7856 | 0.7945 | 0.7900 | 0.9424 |
| 0.0231 | 187.0 | 194293 | 0.3108 | 0.7741 | 0.8134 | 0.7933 | 0.9438 |
| 0.022 | 188.0 | 195332 | 0.3096 | 0.7899 | 0.8091 | 0.7994 | 0.9453 |
| 0.0216 | 189.0 | 196371 | 0.3048 | 0.7848 | 0.7972 | 0.7910 | 0.9448 |
| 0.0192 | 190.0 | 197410 | 0.3289 | 0.7856 | 0.8075 | 0.7964 | 0.9448 |
| 0.0192 | 191.0 | 198449 | 0.3108 | 0.7638 | 0.8046 | 0.7837 | 0.9422 |
| 0.0206 | 192.0 | 199488 | 0.3262 | 0.7896 | 0.7942 | 0.7919 | 0.9425 |
| 0.0227 | 193.0 | 200527 | 0.3360 | 0.7856 | 0.8010 | 0.7932 | 0.9435 |
| 0.0217 | 194.0 | 201566 | 0.3168 | 0.7880 | 0.8026 | 0.7953 | 0.9438 |
| 0.021 | 195.0 | 202605 | 0.3495 | 0.7844 | 0.8057 | 0.7949 | 0.9433 |
| 0.0222 | 196.0 | 203644 | 0.3075 | 0.7746 | 0.8174 | 0.7954 | 0.9463 |
| 0.0227 | 197.0 | 204683 | 0.3246 | 0.7840 | 0.8084 | 0.7960 | 0.9451 |
| 0.0216 | 198.0 | 205722 | 0.3241 | 0.7697 | 0.8111 | 0.7898 | 0.9445 |
| 0.0211 | 199.0 | 206761 | 0.3279 | 0.7825 | 0.8080 | 0.7951 | 0.9449 |
| 0.0213 | 200.0 | 207800 | 0.3228 | 0.7910 | 0.8109 | 0.8008 | 0.9459 |
| 0.0204 | 201.0 | 208839 | 0.3320 | 0.7686 | 0.8163 | 0.7917 | 0.9439 |
| 0.0208 | 202.0 | 209878 | 0.3229 | 0.7917 | 0.8106 | 0.8010 | 0.9443 |
| 0.0206 | 203.0 | 210917 | 0.3157 | 0.7794 | 0.7891 | 0.7843 | 0.9420 |
| 0.0211 | 204.0 | 211956 | 0.3315 | 0.7890 | 0.8169 | 0.8027 | 0.9462 |
| 0.0193 | 205.0 | 212995 | 0.3102 | 0.7733 | 0.8174 | 0.7947 | 0.9442 |
| 0.0221 | 206.0 | 214034 | 0.3311 | 0.7747 | 0.8082 | 0.7911 | 0.9433 |
| 0.0211 | 207.0 | 215073 | 0.3161 | 0.7770 | 0.7974 | 0.7871 | 0.9426 |
| 0.0205 | 208.0 | 216112 | 0.3331 | 0.7938 | 0.8057 | 0.7997 | 0.9448 |
| 0.0201 | 209.0 | 217151 | 0.3379 | 0.7679 | 0.8174 | 0.7919 | 0.9431 |
| 0.0208 | 210.0 | 218190 | 0.3315 | 0.7800 | 0.8039 | 0.7918 | 0.9436 |
| 0.0196 | 211.0 | 219229 | 0.3211 | 0.7835 | 0.8107 | 0.7969 | 0.9447 |
| 0.0199 | 212.0 | 220268 | 0.3343 | 0.7812 | 0.8059 | 0.7934 | 0.9441 |
| 0.0202 | 213.0 | 221307 | 0.3343 | 0.7856 | 0.7956 | 0.7906 | 0.9425 |
| 0.0195 | 214.0 | 222346 | 0.3183 | 0.7894 | 0.7958 | 0.7926 | 0.9435 |
| 0.0205 | 215.0 | 223385 | 0.3334 | 0.7884 | 0.8107 | 0.7994 | 0.9450 |
| 0.0207 | 216.0 | 224424 | 0.3390 | 0.7692 | 0.8197 | 0.7937 | 0.9438 |
| 0.0177 | 217.0 | 225463 | 0.3283 | 0.7791 | 0.7933 | 0.7861 | 0.9422 |
| 0.021 | 218.0 | 226502 | 0.3357 | 0.7864 | 0.8136 | 0.7998 | 0.9454 |
| 0.0225 | 219.0 | 227541 | 0.3169 | 0.7902 | 0.7996 | 0.7948 | 0.9443 |
| 0.0198 | 220.0 | 228580 | 0.3412 | 0.7769 | 0.8100 | 0.7931 | 0.9436 |
| 0.0217 | 221.0 | 229619 | 0.3444 | 0.7921 | 0.8136 | 0.8027 | 0.9452 |
| 0.0176 | 222.0 | 230658 | 0.3332 | 0.7874 | 0.8082 | 0.7977 | 0.9449 |
| 0.0182 | 223.0 | 231697 | 0.3316 | 0.7702 | 0.8156 | 0.7923 | 0.9433 |
| 0.02 | 224.0 | 232736 | 0.3528 | 0.7702 | 0.8041 | 0.7868 | 0.9434 |
| 0.0188 | 225.0 | 233775 | 0.3411 | 0.7706 | 0.7978 | 0.7839 | 0.9424 |
| 0.0201 | 226.0 | 234814 | 0.3330 | 0.7793 | 0.8017 | 0.7903 | 0.9441 |
| 0.0186 | 227.0 | 235853 | 0.3601 | 0.7776 | 0.8010 | 0.7891 | 0.9420 |
| 0.019 | 228.0 | 236892 | 0.3545 | 0.7809 | 0.8093 | 0.7948 | 0.9433 |
| 0.0175 | 229.0 | 237931 | 0.3252 | 0.7711 | 0.8241 | 0.7967 | 0.9445 |
| 0.0171 | 230.0 | 238970 | 0.3310 | 0.7811 | 0.8070 | 0.7938 | 0.9433 |
| 0.0198 | 231.0 | 240009 | 0.3474 | 0.7857 | 0.8100 | 0.7977 | 0.9450 |
| 0.0177 | 232.0 | 241048 | 0.3353 | 0.7794 | 0.7970 | 0.7881 | 0.9424 |
| 0.0194 | 233.0 | 242087 | 0.3768 | 0.7718 | 0.8259 | 0.7979 | 0.9442 |
| 0.0186 | 234.0 | 243126 | 0.3491 | 0.7765 | 0.8172 | 0.7963 | 0.9439 |
| 0.0187 | 235.0 | 244165 | 0.3673 | 0.7800 | 0.7996 | 0.7897 | 0.9423 |
| 0.0163 | 236.0 | 245204 | 0.3397 | 0.7877 | 0.8057 | 0.7966 | 0.9438 |
| 0.0192 | 237.0 | 246243 | 0.3320 | 0.7805 | 0.8118 | 0.7958 | 0.9446 |
| 0.0187 | 238.0 | 247282 | 0.3540 | 0.7575 | 0.8262 | 0.7904 | 0.9418 |
| 0.0179 | 239.0 | 248321 | 0.3610 | 0.7877 | 0.8052 | 0.7963 | 0.9443 |
| 0.0182 | 240.0 | 249360 | 0.3588 | 0.7734 | 0.7947 | 0.7839 | 0.9411 |
| 0.0177 | 241.0 | 250399 | 0.3460 | 0.7781 | 0.8167 | 0.7969 | 0.9437 |
| 0.0173 | 242.0 | 251438 | 0.3598 | 0.7832 | 0.8014 | 0.7922 | 0.9436 |
| 0.0172 | 243.0 | 252477 | 0.3435 | 0.7786 | 0.8246 | 0.8009 | 0.9440 |
| 0.0181 | 244.0 | 253516 | 0.3653 | 0.7908 | 0.8120 | 0.8012 | 0.9451 |
| 0.0195 | 245.0 | 254555 | 0.3303 | 0.7714 | 0.8124 | 0.7913 | 0.9439 |
| 0.0192 | 246.0 | 255594 | 0.3423 | 0.7851 | 0.8160 | 0.8002 | 0.9447 |
| 0.0173 | 247.0 | 256633 | 0.3323 | 0.7780 | 0.8197 | 0.7983 | 0.9446 |
| 0.0197 | 248.0 | 257672 | 0.3343 | 0.7890 | 0.7988 | 0.7939 | 0.9434 |
| 0.017 | 249.0 | 258711 | 0.3493 | 0.7771 | 0.8073 | 0.7919 | 0.9439 |
| 0.0191 | 250.0 | 259750 | 0.3326 | 0.7890 | 0.8062 | 0.7975 | 0.9447 |
| 0.0165 | 251.0 | 260789 | 0.3326 | 0.7688 | 0.7999 | 0.7840 | 0.9416 |
| 0.0176 | 252.0 | 261828 | 0.3403 | 0.7807 | 0.7961 | 0.7883 | 0.9433 |
| 0.0178 | 253.0 | 262867 | 0.3189 | 0.7862 | 0.8091 | 0.7975 | 0.9456 |
| 0.0172 | 254.0 | 263906 | 0.3399 | 0.7932 | 0.8116 | 0.8023 | 0.9466 |
| 0.0163 | 255.0 | 264945 | 0.3478 | 0.7554 | 0.7952 | 0.7748 | 0.9407 |
| 0.0153 | 256.0 | 265984 | 0.3416 | 0.7820 | 0.8131 | 0.7972 | 0.9447 |
| 0.0187 | 257.0 | 267023 | 0.3300 | 0.7790 | 0.8100 | 0.7942 | 0.9442 |
| 0.0185 | 258.0 | 268062 | 0.3292 | 0.7854 | 0.7981 | 0.7917 | 0.9445 |
| 0.0177 | 259.0 | 269101 | 0.3468 | 0.7858 | 0.8179 | 0.8016 | 0.9451 |
| 0.0155 | 260.0 | 270140 | 0.3416 | 0.7596 | 0.7990 | 0.7788 | 0.9415 |
| 0.0176 | 261.0 | 271179 | 0.3435 | 0.7642 | 0.8298 | 0.7956 | 0.9435 |
| 0.0172 | 262.0 | 272218 | 0.3448 | 0.7817 | 0.7853 | 0.7835 | 0.9406 |
| 0.0161 | 263.0 | 273257 | 0.3306 | 0.7703 | 0.8075 | 0.7885 | 0.9418 |
| 0.0176 | 264.0 | 274296 | 0.3641 | 0.7767 | 0.8017 | 0.7890 | 0.9407 |
| 0.0172 | 265.0 | 275335 | 0.3492 | 0.7881 | 0.8098 | 0.7988 | 0.9436 |
| 0.0173 | 266.0 | 276374 | 0.3297 | 0.7971 | 0.8008 | 0.7990 | 0.9446 |
| 0.0177 | 267.0 | 277413 | 0.3643 | 0.7817 | 0.8156 | 0.7983 | 0.9442 |
| 0.016 | 268.0 | 278452 | 0.3521 | 0.7812 | 0.8142 | 0.7974 | 0.9446 |
| 0.0155 | 269.0 | 279491 | 0.3529 | 0.7717 | 0.8174 | 0.7939 | 0.9449 |
| 0.0177 | 270.0 | 280530 | 0.3479 | 0.7833 | 0.7949 | 0.7891 | 0.9424 |
| 0.0171 | 271.0 | 281569 | 0.3481 | 0.7864 | 0.8088 | 0.7974 | 0.9441 |
| 0.0176 | 272.0 | 282608 | 0.3356 | 0.7820 | 0.8152 | 0.7983 | 0.9442 |
| 0.017 | 273.0 | 283647 | 0.3488 | 0.7762 | 0.8044 | 0.7901 | 0.9432 |
| 0.0188 | 274.0 | 284686 | 0.3473 | 0.7907 | 0.8033 | 0.7970 | 0.9436 |
| 0.0162 | 275.0 | 285725 | 0.3469 | 0.7890 | 0.8187 | 0.8035 | 0.9453 |
| 0.0159 | 276.0 | 286764 | 0.3306 | 0.7948 | 0.8215 | 0.8079 | 0.9465 |
| 0.0147 | 277.0 | 287803 | 0.3399 | 0.7970 | 0.8100 | 0.8034 | 0.9457 |
| 0.0155 | 278.0 | 288842 | 0.3403 | 0.7846 | 0.8037 | 0.7941 | 0.9440 |
| 0.0141 | 279.0 | 289881 | 0.3460 | 0.7760 | 0.8046 | 0.7900 | 0.9427 |
| 0.0146 | 280.0 | 290920 | 0.3253 | 0.7810 | 0.8023 | 0.7915 | 0.9439 |
| 0.0156 | 281.0 | 291959 | 0.3400 | 0.7820 | 0.8080 | 0.7948 | 0.9439 |
| 0.0141 | 282.0 | 292998 | 0.3019 | 0.7944 | 0.7994 | 0.7969 | 0.9444 |
| 0.0157 | 283.0 | 294037 | 0.3610 | 0.7882 | 0.8136 | 0.8007 | 0.9454 |
| 0.0176 | 284.0 | 295076 | 0.3498 | 0.7767 | 0.8023 | 0.7893 | 0.9430 |
| 0.015 | 285.0 | 296115 | 0.3594 | 0.7722 | 0.7943 | 0.7831 | 0.9426 |
| 0.0161 | 286.0 | 297154 | 0.3613 | 0.7795 | 0.8138 | 0.7963 | 0.9438 |
| 0.0157 | 287.0 | 298193 | 0.3481 | 0.7824 | 0.7940 | 0.7882 | 0.9409 |
| 0.0161 | 288.0 | 299232 | 0.3201 | 0.7725 | 0.8134 | 0.7925 | 0.9446 |
| 0.0141 | 289.0 | 300271 | 0.3533 | 0.7821 | 0.8106 | 0.7961 | 0.9444 |
| 0.0149 | 290.0 | 301310 | 0.3384 | 0.7746 | 0.8138 | 0.7937 | 0.9444 |
| 0.0145 | 291.0 | 302349 | 0.3247 | 0.7836 | 0.8028 | 0.7931 | 0.9435 |
| 0.0155 | 292.0 | 303388 | 0.3420 | 0.7762 | 0.8093 | 0.7924 | 0.9444 |
| 0.0147 | 293.0 | 304427 | 0.3206 | 0.7821 | 0.8183 | 0.7998 | 0.9460 |
| 0.014 | 294.0 | 305466 | 0.3537 | 0.7869 | 0.8158 | 0.8011 | 0.9447 |
| 0.0163 | 295.0 | 306505 | 0.3416 | 0.7741 | 0.8111 | 0.7922 | 0.9443 |
| 0.0176 | 296.0 | 307544 | 0.3449 | 0.7855 | 0.8120 | 0.7985 | 0.9440 |
| 0.0168 | 297.0 | 308583 | 0.3476 | 0.7852 | 0.8170 | 0.8008 | 0.9447 |
| 0.0149 | 298.0 | 309622 | 0.3383 | 0.7828 | 0.8086 | 0.7955 | 0.9427 |
| 0.0162 | 299.0 | 310661 | 0.3481 | 0.7847 | 0.8059 | 0.7951 | 0.9432 |
| 0.0165 | 300.0 | 311700 | 0.3431 | 0.7828 | 0.8163 | 0.7992 | 0.9451 |
| 0.0146 | 301.0 | 312739 | 0.3549 | 0.7868 | 0.7981 | 0.7924 | 0.9416 |
| 0.0156 | 302.0 | 313778 | 0.3488 | 0.7646 | 0.8147 | 0.7888 | 0.9430 |
| 0.0133 | 303.0 | 314817 | 0.3357 | 0.7756 | 0.8122 | 0.7935 | 0.9450 |
| 0.0147 | 304.0 | 315856 | 0.3498 | 0.7757 | 0.8071 | 0.7911 | 0.9430 |
| 0.013 | 305.0 | 316895 | 0.3488 | 0.7726 | 0.8106 | 0.7911 | 0.9443 |
| 0.0137 | 306.0 | 317934 | 0.3634 | 0.7831 | 0.7895 | 0.7863 | 0.9425 |
| 0.0148 | 307.0 | 318973 | 0.3646 | 0.7816 | 0.8179 | 0.7994 | 0.9447 |
| 0.0151 | 308.0 | 320012 | 0.3213 | 0.7774 | 0.8061 | 0.7914 | 0.9427 |
| 0.0151 | 309.0 | 321051 | 0.3735 | 0.7774 | 0.8149 | 0.7957 | 0.9434 |
| 0.0165 | 310.0 | 322090 | 0.3389 | 0.7898 | 0.8006 | 0.7952 | 0.9440 |
| 0.0128 | 311.0 | 323129 | 0.3577 | 0.7760 | 0.8228 | 0.7987 | 0.9447 |
| 0.0139 | 312.0 | 324168 | 0.3309 | 0.7826 | 0.8070 | 0.7946 | 0.9455 |
| 0.0144 | 313.0 | 325207 | 0.3494 | 0.7748 | 0.8147 | 0.7942 | 0.9436 |
| 0.0139 | 314.0 | 326246 | 0.3397 | 0.7939 | 0.8014 | 0.7976 | 0.9439 |
| 0.0149 | 315.0 | 327285 | 0.3345 | 0.7826 | 0.8053 | 0.7938 | 0.9448 |
| 0.0152 | 316.0 | 328324 | 0.3359 | 0.7860 | 0.8223 | 0.8037 | 0.9454 |
| 0.0137 | 317.0 | 329363 | 0.3734 | 0.7810 | 0.8158 | 0.7980 | 0.9433 |
| 0.0146 | 318.0 | 330402 | 0.3523 | 0.7793 | 0.8052 | 0.7920 | 0.9426 |
| 0.0129 | 319.0 | 331441 | 0.3644 | 0.7899 | 0.8050 | 0.7974 | 0.9446 |
| 0.0117 | 320.0 | 332480 | 0.3527 | 0.7838 | 0.8055 | 0.7945 | 0.9429 |
| 0.0166 | 321.0 | 333519 | 0.3529 | 0.7856 | 0.8237 | 0.8042 | 0.9452 |
| 0.0146 | 322.0 | 334558 | 0.3277 | 0.7813 | 0.8147 | 0.7977 | 0.9450 |
| 0.0148 | 323.0 | 335597 | 0.3370 | 0.7861 | 0.8161 | 0.8008 | 0.9450 |
| 0.0139 | 324.0 | 336636 | 0.3834 | 0.7909 | 0.8052 | 0.7980 | 0.9434 |
| 0.0133 | 325.0 | 337675 | 0.3933 | 0.7801 | 0.7972 | 0.7886 | 0.9415 |
| 0.0138 | 326.0 | 338714 | 0.3436 | 0.7860 | 0.7925 | 0.7893 | 0.9428 |
| 0.0131 | 327.0 | 339753 | 0.3518 | 0.7872 | 0.8116 | 0.7993 | 0.9450 |
| 0.0142 | 328.0 | 340792 | 0.3439 | 0.7921 | 0.8039 | 0.7979 | 0.9444 |
| 0.0141 | 329.0 | 341831 | 0.3545 | 0.7934 | 0.8154 | 0.8043 | 0.9451 |
| 0.0119 | 330.0 | 342870 | 0.3294 | 0.7949 | 0.8048 | 0.7998 | 0.9449 |
| 0.0141 | 331.0 | 343909 | 0.3390 | 0.7886 | 0.8050 | 0.7967 | 0.9444 |
| 0.0128 | 332.0 | 344948 | 0.3819 | 0.7929 | 0.8046 | 0.7987 | 0.9448 |
| 0.0142 | 333.0 | 345987 | 0.3477 | 0.7907 | 0.8028 | 0.7967 | 0.9440 |
| 0.0133 | 334.0 | 347026 | 0.3668 | 0.7762 | 0.8120 | 0.7937 | 0.9434 |
| 0.0126 | 335.0 | 348065 | 0.3437 | 0.7859 | 0.8185 | 0.8019 | 0.9450 |
| 0.0133 | 336.0 | 349104 | 0.3728 | 0.7868 | 0.8122 | 0.7993 | 0.9447 |
| 0.0121 | 337.0 | 350143 | 0.3867 | 0.7818 | 0.8109 | 0.7961 | 0.9438 |
| 0.0146 | 338.0 | 351182 | 0.3942 | 0.7798 | 0.8059 | 0.7926 | 0.9408 |
| 0.014 | 339.0 | 352221 | 0.3535 | 0.7735 | 0.8033 | 0.7882 | 0.9427 |
| 0.0122 | 340.0 | 353260 | 0.3543 | 0.7733 | 0.7994 | 0.7862 | 0.9423 |
| 0.0115 | 341.0 | 354299 | 0.3537 | 0.7745 | 0.8021 | 0.7880 | 0.9418 |
| 0.0128 | 342.0 | 355338 | 0.3637 | 0.7797 | 0.8077 | 0.7935 | 0.9435 |
| 0.0115 | 343.0 | 356377 | 0.3860 | 0.7840 | 0.8106 | 0.7971 | 0.9428 |
| 0.0122 | 344.0 | 357416 | 0.3592 | 0.7791 | 0.8079 | 0.7932 | 0.9436 |
| 0.0112 | 345.0 | 358455 | 0.3458 | 0.7855 | 0.8001 | 0.7928 | 0.9437 |
| 0.012 | 346.0 | 359494 | 0.3819 | 0.7884 | 0.8010 | 0.7946 | 0.9434 |
| 0.0126 | 347.0 | 360533 | 0.3786 | 0.7716 | 0.8068 | 0.7888 | 0.9432 |
| 0.013 | 348.0 | 361572 | 0.3623 | 0.7854 | 0.8106 | 0.7978 | 0.9439 |
| 0.0152 | 349.0 | 362611 | 0.3389 | 0.7881 | 0.8102 | 0.7990 | 0.9450 |
| 0.0112 | 350.0 | 363650 | 0.3541 | 0.7843 | 0.8098 | 0.7968 | 0.9446 |
| 0.0124 | 351.0 | 364689 | 0.3576 | 0.7792 | 0.8104 | 0.7945 | 0.9431 |
| 0.0124 | 352.0 | 365728 | 0.3694 | 0.7838 | 0.8152 | 0.7992 | 0.9438 |
| 0.013 | 353.0 | 366767 | 0.3461 | 0.7882 | 0.8080 | 0.7980 | 0.9433 |
| 0.014 | 354.0 | 367806 | 0.3824 | 0.7734 | 0.7990 | 0.7860 | 0.9410 |
| 0.0111 | 355.0 | 368845 | 0.3578 | 0.7904 | 0.8014 | 0.7959 | 0.9435 |
| 0.012 | 356.0 | 369884 | 0.3593 | 0.7873 | 0.8068 | 0.7969 | 0.9446 |
| 0.0124 | 357.0 | 370923 | 0.3673 | 0.7915 | 0.8048 | 0.7981 | 0.9426 |
| 0.0126 | 358.0 | 371962 | 0.3591 | 0.7842 | 0.8142 | 0.7989 | 0.9444 |
| 0.0132 | 359.0 | 373001 | 0.3682 | 0.7844 | 0.8079 | 0.7960 | 0.9440 |
| 0.0132 | 360.0 | 374040 | 0.3725 | 0.7854 | 0.8140 | 0.7994 | 0.9439 |
| 0.0128 | 361.0 | 375079 | 0.3438 | 0.7909 | 0.8113 | 0.8010 | 0.9445 |
| 0.0121 | 362.0 | 376118 | 0.3889 | 0.7860 | 0.8021 | 0.7939 | 0.9421 |
| 0.0127 | 363.0 | 377157 | 0.3702 | 0.7817 | 0.8158 | 0.7984 | 0.9441 |
| 0.0124 | 364.0 | 378196 | 0.3529 | 0.7805 | 0.8075 | 0.7938 | 0.9435 |
| 0.0128 | 365.0 | 379235 | 0.3773 | 0.7850 | 0.8125 | 0.7985 | 0.9429 |
| 0.0117 | 366.0 | 380274 | 0.3635 | 0.7855 | 0.8205 | 0.8026 | 0.9454 |
| 0.0114 | 367.0 | 381313 | 0.3499 | 0.7825 | 0.8133 | 0.7976 | 0.9435 |
| 0.0111 | 368.0 | 382352 | 0.3495 | 0.7847 | 0.8156 | 0.7998 | 0.9442 |
| 0.0121 | 369.0 | 383391 | 0.3526 | 0.7844 | 0.8080 | 0.7961 | 0.9436 |
| 0.0115 | 370.0 | 384430 | 0.3585 | 0.7931 | 0.8113 | 0.8021 | 0.9444 |
| 0.0126 | 371.0 | 385469 | 0.3643 | 0.7757 | 0.8197 | 0.7971 | 0.9440 |
| 0.0132 | 372.0 | 386508 | 0.3442 | 0.7939 | 0.8053 | 0.7996 | 0.9435 |
| 0.0127 | 373.0 | 387547 | 0.3762 | 0.7825 | 0.8152 | 0.7986 | 0.9434 |
| 0.0132 | 374.0 | 388586 | 0.3675 | 0.7711 | 0.7994 | 0.7850 | 0.9409 |
| 0.0119 | 375.0 | 389625 | 0.3624 | 0.7762 | 0.8170 | 0.7961 | 0.9423 |
| 0.0125 | 376.0 | 390664 | 0.3796 | 0.7930 | 0.7920 | 0.7925 | 0.9430 |
| 0.0107 | 377.0 | 391703 | 0.3629 | 0.7840 | 0.8203 | 0.8017 | 0.9454 |
| 0.0129 | 378.0 | 392742 | 0.3430 | 0.7928 | 0.8122 | 0.8023 | 0.9440 |
| 0.011 | 379.0 | 393781 | 0.3596 | 0.7805 | 0.8115 | 0.7957 | 0.9431 |
| 0.0108 | 380.0 | 394820 | 0.3668 | 0.7815 | 0.8102 | 0.7956 | 0.9425 |
| 0.01 | 381.0 | 395859 | 0.3524 | 0.7755 | 0.8048 | 0.7899 | 0.9422 |
| 0.0111 | 382.0 | 396898 | 0.3805 | 0.7786 | 0.8089 | 0.7935 | 0.9432 |
| 0.0098 | 383.0 | 397937 | 0.3719 | 0.7931 | 0.8082 | 0.8006 | 0.9438 |
| 0.0113 | 384.0 | 398976 | 0.3659 | 0.7825 | 0.8071 | 0.7946 | 0.9433 |
| 0.0137 | 385.0 | 400015 | 0.3710 | 0.7872 | 0.8313 | 0.8086 | 0.9452 |
| 0.0125 | 386.0 | 401054 | 0.3621 | 0.7963 | 0.8012 | 0.7987 | 0.9428 |
| 0.011 | 387.0 | 402093 | 0.3650 | 0.7868 | 0.8001 | 0.7934 | 0.9430 |
| 0.0116 | 388.0 | 403132 | 0.3744 | 0.7884 | 0.8224 | 0.8050 | 0.9450 |
| 0.0103 | 389.0 | 404171 | 0.3562 | 0.7884 | 0.8253 | 0.8064 | 0.9454 |
| 0.012 | 390.0 | 405210 | 0.3662 | 0.7794 | 0.8088 | 0.7938 | 0.9425 |
| 0.0098 | 391.0 | 406249 | 0.3516 | 0.7868 | 0.8107 | 0.7986 | 0.9451 |
| 0.0126 | 392.0 | 407288 | 0.3481 | 0.7907 | 0.8015 | 0.7961 | 0.9432 |
| 0.0102 | 393.0 | 408327 | 0.3665 | 0.7868 | 0.8179 | 0.8020 | 0.9448 |
| 0.0113 | 394.0 | 409366 | 0.3466 | 0.7860 | 0.8151 | 0.8003 | 0.9450 |
| 0.0103 | 395.0 | 410405 | 0.3706 | 0.7888 | 0.8210 | 0.8046 | 0.9453 |
| 0.0114 | 396.0 | 411444 | 0.3326 | 0.7920 | 0.8181 | 0.8049 | 0.9454 |
| 0.0101 | 397.0 | 412483 | 0.3605 | 0.7839 | 0.8089 | 0.7962 | 0.9433 |
| 0.0114 | 398.0 | 413522 | 0.3731 | 0.7828 | 0.8091 | 0.7957 | 0.9433 |
| 0.0102 | 399.0 | 414561 | 0.3839 | 0.7932 | 0.8118 | 0.8024 | 0.9441 |
| 0.0101 | 400.0 | 415600 | 0.3379 | 0.7954 | 0.8080 | 0.8017 | 0.9445 |
| 0.0113 | 401.0 | 416639 | 0.3638 | 0.7820 | 0.8089 | 0.7953 | 0.9439 |
| 0.011 | 402.0 | 417678 | 0.3745 | 0.7916 | 0.8138 | 0.8025 | 0.9455 |
| 0.0095 | 403.0 | 418717 | 0.3398 | 0.7912 | 0.7983 | 0.7947 | 0.9436 |
| 0.0112 | 404.0 | 419756 | 0.3511 | 0.7889 | 0.8075 | 0.7981 | 0.9438 |
| 0.0105 | 405.0 | 420795 | 0.3366 | 0.7866 | 0.8118 | 0.7990 | 0.9439 |
| 0.0105 | 406.0 | 421834 | 0.3776 | 0.7900 | 0.8028 | 0.7964 | 0.9426 |
| 0.0095 | 407.0 | 422873 | 0.3801 | 0.7907 | 0.8089 | 0.7997 | 0.9428 |
| 0.0098 | 408.0 | 423912 | 0.3896 | 0.7903 | 0.8048 | 0.7975 | 0.9433 |
| 0.0119 | 409.0 | 424951 | 0.3674 | 0.7885 | 0.8206 | 0.8043 | 0.9456 |
| 0.0098 | 410.0 | 425990 | 0.3893 | 0.7953 | 0.8208 | 0.8079 | 0.9473 |
| 0.011 | 411.0 | 427029 | 0.3585 | 0.7917 | 0.8082 | 0.7999 | 0.9439 |
| 0.0102 | 412.0 | 428068 | 0.3466 | 0.7814 | 0.8100 | 0.7955 | 0.9440 |
| 0.0095 | 413.0 | 429107 | 0.3510 | 0.7851 | 0.8044 | 0.7946 | 0.9443 |
| 0.0106 | 414.0 | 430146 | 0.3690 | 0.7942 | 0.7979 | 0.7961 | 0.9434 |
| 0.0101 | 415.0 | 431185 | 0.3748 | 0.7882 | 0.8017 | 0.7949 | 0.9437 |
| 0.01 | 416.0 | 432224 | 0.3827 | 0.7975 | 0.8107 | 0.8041 | 0.9451 |
| 0.0098 | 417.0 | 433263 | 0.3955 | 0.7866 | 0.8163 | 0.8012 | 0.9439 |
| 0.0097 | 418.0 | 434302 | 0.3802 | 0.7813 | 0.8102 | 0.7955 | 0.9436 |
| 0.0113 | 419.0 | 435341 | 0.3745 | 0.7932 | 0.8122 | 0.8026 | 0.9440 |
| 0.01 | 420.0 | 436380 | 0.3807 | 0.7820 | 0.8088 | 0.7951 | 0.9427 |
| 0.01 | 421.0 | 437419 | 0.3646 | 0.7856 | 0.8118 | 0.7985 | 0.9447 |
| 0.0108 | 422.0 | 438458 | 0.3762 | 0.7940 | 0.8129 | 0.8033 | 0.9447 |
| 0.0083 | 423.0 | 439497 | 0.3698 | 0.7910 | 0.8106 | 0.8007 | 0.9444 |
| 0.0111 | 424.0 | 440536 | 0.3601 | 0.7886 | 0.8127 | 0.8005 | 0.9442 |
| 0.0115 | 425.0 | 441575 | 0.3628 | 0.7875 | 0.8115 | 0.7993 | 0.9440 |
| 0.0094 | 426.0 | 442614 | 0.3798 | 0.7921 | 0.8095 | 0.8007 | 0.9459 |
| 0.0102 | 427.0 | 443653 | 0.3651 | 0.7901 | 0.8093 | 0.7996 | 0.9447 |
| 0.0103 | 428.0 | 444692 | 0.3981 | 0.7850 | 0.8221 | 0.8031 | 0.9454 |
| 0.0087 | 429.0 | 445731 | 0.3641 | 0.7731 | 0.8077 | 0.7900 | 0.9446 |
| 0.01 | 430.0 | 446770 | 0.3666 | 0.7934 | 0.8138 | 0.8034 | 0.9449 |
| 0.0111 | 431.0 | 447809 | 0.3671 | 0.7868 | 0.8134 | 0.7999 | 0.9448 |
| 0.0091 | 432.0 | 448848 | 0.3768 | 0.7923 | 0.8133 | 0.8026 | 0.9457 |
| 0.0087 | 433.0 | 449887 | 0.3878 | 0.7917 | 0.8089 | 0.8002 | 0.9453 |
| 0.009 | 434.0 | 450926 | 0.3549 | 0.7906 | 0.8174 | 0.8038 | 0.9461 |
| 0.01 | 435.0 | 451965 | 0.3683 | 0.7902 | 0.7985 | 0.7943 | 0.9442 |
| 0.0097 | 436.0 | 453004 | 0.3642 | 0.7821 | 0.8158 | 0.7986 | 0.9444 |
| 0.0097 | 437.0 | 454043 | 0.3925 | 0.7826 | 0.8151 | 0.7985 | 0.9461 |
| 0.0095 | 438.0 | 455082 | 0.3749 | 0.7777 | 0.8003 | 0.7889 | 0.9432 |
| 0.0109 | 439.0 | 456121 | 0.3612 | 0.7906 | 0.8116 | 0.8010 | 0.9451 |
| 0.0088 | 440.0 | 457160 | 0.3803 | 0.7886 | 0.8102 | 0.7993 | 0.9449 |
| 0.0079 | 441.0 | 458199 | 0.3566 | 0.7761 | 0.8167 | 0.7959 | 0.9444 |
| 0.0086 | 442.0 | 459238 | 0.3849 | 0.7855 | 0.8143 | 0.7996 | 0.9446 |
| 0.0094 | 443.0 | 460277 | 0.3864 | 0.7756 | 0.8197 | 0.7971 | 0.9443 |
| 0.0087 | 444.0 | 461316 | 0.3944 | 0.7737 | 0.8133 | 0.7930 | 0.9439 |
| 0.0104 | 445.0 | 462355 | 0.3872 | 0.7777 | 0.8176 | 0.7971 | 0.9443 |
| 0.0086 | 446.0 | 463394 | 0.3769 | 0.7842 | 0.8125 | 0.7981 | 0.9449 |
| 0.0096 | 447.0 | 464433 | 0.3585 | 0.7834 | 0.8147 | 0.7987 | 0.9447 |
| 0.009 | 448.0 | 465472 | 0.3583 | 0.7795 | 0.8097 | 0.7943 | 0.9438 |
| 0.01 | 449.0 | 466511 | 0.3542 | 0.7800 | 0.8215 | 0.8002 | 0.9458 |
| 0.0091 | 450.0 | 467550 | 0.3506 | 0.7868 | 0.8109 | 0.7987 | 0.9457 |
| 0.01 | 451.0 | 468589 | 0.3771 | 0.7902 | 0.8235 | 0.8065 | 0.9465 |
| 0.0091 | 452.0 | 469628 | 0.3702 | 0.7889 | 0.8084 | 0.7985 | 0.9436 |
| 0.009 | 453.0 | 470667 | 0.3680 | 0.7855 | 0.8215 | 0.8031 | 0.9454 |
| 0.0083 | 454.0 | 471706 | 0.4019 | 0.7810 | 0.8079 | 0.7942 | 0.9443 |
| 0.0094 | 455.0 | 472745 | 0.3821 | 0.7742 | 0.8138 | 0.7935 | 0.9428 |
| 0.0092 | 456.0 | 473784 | 0.3624 | 0.7956 | 0.8181 | 0.8067 | 0.9461 |
| 0.0099 | 457.0 | 474823 | 0.3863 | 0.7852 | 0.8248 | 0.8045 | 0.9460 |
| 0.0085 | 458.0 | 475862 | 0.3794 | 0.7911 | 0.8161 | 0.8034 | 0.9446 |
| 0.008 | 459.0 | 476901 | 0.3801 | 0.7951 | 0.8196 | 0.8071 | 0.9467 |
| 0.0086 | 460.0 | 477940 | 0.3926 | 0.7900 | 0.8156 | 0.8026 | 0.9444 |
| 0.0088 | 461.0 | 478979 | 0.3718 | 0.8007 | 0.8296 | 0.8149 | 0.9465 |
| 0.009 | 462.0 | 480018 | 0.3865 | 0.7867 | 0.8134 | 0.7998 | 0.9443 |
| 0.0106 | 463.0 | 481057 | 0.3809 | 0.7976 | 0.8145 | 0.8060 | 0.9455 |
| 0.0101 | 464.0 | 482096 | 0.3658 | 0.8005 | 0.8131 | 0.8068 | 0.9458 |
| 0.0087 | 465.0 | 483135 | 0.3836 | 0.7959 | 0.8131 | 0.8044 | 0.9454 |
| 0.0086 | 466.0 | 484174 | 0.4016 | 0.7920 | 0.8124 | 0.8020 | 0.9450 |
| 0.0088 | 467.0 | 485213 | 0.3742 | 0.7833 | 0.8077 | 0.7953 | 0.9428 |
| 0.0091 | 468.0 | 486252 | 0.3755 | 0.7945 | 0.8172 | 0.8057 | 0.9461 |
| 0.0082 | 469.0 | 487291 | 0.4018 | 0.7814 | 0.8226 | 0.8015 | 0.9448 |
| 0.0078 | 470.0 | 488330 | 0.3838 | 0.7958 | 0.8156 | 0.8056 | 0.9461 |
| 0.0093 | 471.0 | 489369 | 0.3638 | 0.7861 | 0.8106 | 0.7981 | 0.9447 |
| 0.0075 | 472.0 | 490408 | 0.3524 | 0.7899 | 0.8179 | 0.8037 | 0.9456 |
| 0.0081 | 473.0 | 491447 | 0.3804 | 0.7842 | 0.8084 | 0.7961 | 0.9431 |
| 0.0081 | 474.0 | 492486 | 0.3866 | 0.7838 | 0.8082 | 0.7958 | 0.9431 |
| 0.0082 | 475.0 | 493525 | 0.3907 | 0.7859 | 0.8156 | 0.8005 | 0.9442 |
| 0.0074 | 476.0 | 494564 | 0.3899 | 0.7851 | 0.7996 | 0.7923 | 0.9416 |
| 0.009 | 477.0 | 495603 | 0.3617 | 0.7841 | 0.7992 | 0.7916 | 0.9429 |
| 0.0075 | 478.0 | 496642 | 0.3865 | 0.7943 | 0.8026 | 0.7985 | 0.9433 |
| 0.0073 | 479.0 | 497681 | 0.3827 | 0.7987 | 0.8131 | 0.8058 | 0.9450 |
| 0.009 | 480.0 | 498720 | 0.3815 | 0.7822 | 0.8143 | 0.7980 | 0.9438 |
| 0.0081 | 481.0 | 499759 | 0.3492 | 0.7899 | 0.8140 | 0.8018 | 0.9442 |
| 0.0075 | 482.0 | 500798 | 0.3606 | 0.7946 | 0.8233 | 0.8087 | 0.9463 |
| 0.0072 | 483.0 | 501837 | 0.3900 | 0.7900 | 0.8183 | 0.8039 | 0.9458 |
| 0.0096 | 484.0 | 502876 | 0.3813 | 0.7872 | 0.8194 | 0.8030 | 0.9453 |
| 0.0081 | 485.0 | 503915 | 0.4023 | 0.7865 | 0.8093 | 0.7977 | 0.9444 |
| 0.008 | 486.0 | 504954 | 0.4004 | 0.7834 | 0.8143 | 0.7986 | 0.9440 |
| 0.0069 | 487.0 | 505993 | 0.4039 | 0.7891 | 0.8111 | 0.7999 | 0.9435 |
| 0.0093 | 488.0 | 507032 | 0.3894 | 0.7901 | 0.8140 | 0.8018 | 0.9447 |
| 0.0066 | 489.0 | 508071 | 0.3817 | 0.7891 | 0.8134 | 0.8011 | 0.9447 |
| 0.0078 | 490.0 | 509110 | 0.3806 | 0.7914 | 0.8053 | 0.7983 | 0.9436 |
| 0.0089 | 491.0 | 510149 | 0.3932 | 0.7908 | 0.8044 | 0.7975 | 0.9439 |
| 0.0085 | 492.0 | 511188 | 0.4089 | 0.7951 | 0.8098 | 0.8024 | 0.9452 |
| 0.0081 | 493.0 | 512227 | 0.4139 | 0.7805 | 0.8041 | 0.7921 | 0.9431 |
| 0.0081 | 494.0 | 513266 | 0.3761 | 0.7890 | 0.8179 | 0.8032 | 0.9447 |
| 0.0078 | 495.0 | 514305 | 0.4036 | 0.7930 | 0.8235 | 0.8080 | 0.9459 |
| 0.0083 | 496.0 | 515344 | 0.3714 | 0.7950 | 0.8158 | 0.8053 | 0.9456 |
| 0.0085 | 497.0 | 516383 | 0.3776 | 0.7934 | 0.8015 | 0.7975 | 0.9439 |
| 0.0089 | 498.0 | 517422 | 0.3969 | 0.7872 | 0.8142 | 0.8005 | 0.9441 |
| 0.0079 | 499.0 | 518461 | 0.3750 | 0.7951 | 0.8122 | 0.8036 | 0.9454 |
| 0.0068 | 500.0 | 519500 | 0.3919 | 0.7976 | 0.8102 | 0.8038 | 0.9457 |
| 0.0088 | 501.0 | 520539 | 0.3727 | 0.7798 | 0.8226 | 0.8006 | 0.9433 |
| 0.0083 | 502.0 | 521578 | 0.3907 | 0.7837 | 0.8201 | 0.8015 | 0.9454 |
| 0.0088 | 503.0 | 522617 | 0.3784 | 0.7794 | 0.8032 | 0.7911 | 0.9436 |
| 0.0076 | 504.0 | 523656 | 0.3909 | 0.7845 | 0.8147 | 0.7993 | 0.9442 |
| 0.0074 | 505.0 | 524695 | 0.3740 | 0.7890 | 0.8190 | 0.8037 | 0.9452 |
| 0.0067 | 506.0 | 525734 | 0.3796 | 0.7782 | 0.8212 | 0.7991 | 0.9447 |
| 0.0062 | 507.0 | 526773 | 0.3793 | 0.7808 | 0.8102 | 0.7952 | 0.9448 |
| 0.0079 | 508.0 | 527812 | 0.3933 | 0.7998 | 0.8129 | 0.8063 | 0.9464 |
| 0.0074 | 509.0 | 528851 | 0.4202 | 0.7873 | 0.8201 | 0.8034 | 0.9459 |
| 0.0074 | 510.0 | 529890 | 0.3957 | 0.7858 | 0.8206 | 0.8029 | 0.9457 |
| 0.0071 | 511.0 | 530929 | 0.4042 | 0.7898 | 0.8023 | 0.7960 | 0.9446 |
| 0.0068 | 512.0 | 531968 | 0.3984 | 0.7860 | 0.8201 | 0.8027 | 0.9454 |
| 0.0075 | 513.0 | 533007 | 0.4054 | 0.7850 | 0.8062 | 0.7955 | 0.9433 |
| 0.0075 | 514.0 | 534046 | 0.4435 | 0.7976 | 0.7988 | 0.7982 | 0.9435 |
| 0.0081 | 515.0 | 535085 | 0.3894 | 0.7919 | 0.8244 | 0.8078 | 0.9457 |
| 0.0083 | 516.0 | 536124 | 0.3802 | 0.7947 | 0.8136 | 0.8041 | 0.9448 |
| 0.0065 | 517.0 | 537163 | 0.3587 | 0.7950 | 0.8226 | 0.8086 | 0.9472 |
| 0.0072 | 518.0 | 538202 | 0.3751 | 0.8074 | 0.8199 | 0.8136 | 0.9476 |
| 0.0079 | 519.0 | 539241 | 0.3975 | 0.7948 | 0.8167 | 0.8056 | 0.9464 |
| 0.0063 | 520.0 | 540280 | 0.4078 | 0.7831 | 0.8187 | 0.8005 | 0.9454 |
| 0.0072 | 521.0 | 541319 | 0.3884 | 0.7939 | 0.8246 | 0.8089 | 0.9459 |
| 0.007 | 522.0 | 542358 | 0.4058 | 0.7898 | 0.8044 | 0.7970 | 0.9437 |
| 0.0063 | 523.0 | 543397 | 0.3937 | 0.7906 | 0.8133 | 0.8018 | 0.9446 |
| 0.0072 | 524.0 | 544436 | 0.4096 | 0.7986 | 0.7992 | 0.7989 | 0.9444 |
| 0.0063 | 525.0 | 545475 | 0.4032 | 0.7921 | 0.8233 | 0.8074 | 0.9464 |
| 0.0065 | 526.0 | 546514 | 0.3913 | 0.7909 | 0.8050 | 0.7979 | 0.9434 |
| 0.0079 | 527.0 | 547553 | 0.4097 | 0.7974 | 0.8008 | 0.7991 | 0.9431 |
| 0.0072 | 528.0 | 548592 | 0.4014 | 0.7891 | 0.8088 | 0.7988 | 0.9447 |
| 0.0078 | 529.0 | 549631 | 0.3901 | 0.7880 | 0.8219 | 0.8046 | 0.9450 |
| 0.0065 | 530.0 | 550670 | 0.3945 | 0.7979 | 0.8064 | 0.8021 | 0.9445 |
| 0.0054 | 531.0 | 551709 | 0.4075 | 0.7874 | 0.8138 | 0.8004 | 0.9450 |
| 0.0069 | 532.0 | 552748 | 0.4054 | 0.7796 | 0.8095 | 0.7942 | 0.9432 |
| 0.0054 | 533.0 | 553787 | 0.3829 | 0.7970 | 0.8046 | 0.8008 | 0.9437 |
| 0.0069 | 534.0 | 554826 | 0.3878 | 0.7947 | 0.8066 | 0.8006 | 0.9431 |
| 0.0067 | 535.0 | 555865 | 0.3710 | 0.7950 | 0.8071 | 0.8010 | 0.9448 |
| 0.0078 | 536.0 | 556904 | 0.3909 | 0.7935 | 0.8026 | 0.7980 | 0.9441 |
| 0.007 | 537.0 | 557943 | 0.4149 | 0.7897 | 0.8151 | 0.8022 | 0.9459 |
| 0.0062 | 538.0 | 558982 | 0.4012 | 0.7965 | 0.8093 | 0.8029 | 0.9448 |
| 0.0065 | 539.0 | 560021 | 0.4038 | 0.7918 | 0.8102 | 0.8009 | 0.9433 |
| 0.0074 | 540.0 | 561060 | 0.3636 | 0.7828 | 0.8118 | 0.7970 | 0.9436 |
| 0.0067 | 541.0 | 562099 | 0.4026 | 0.7863 | 0.8151 | 0.8004 | 0.9440 |
| 0.0067 | 542.0 | 563138 | 0.4191 | 0.7772 | 0.8169 | 0.7966 | 0.9436 |
| 0.0062 | 543.0 | 564177 | 0.4094 | 0.7870 | 0.8116 | 0.7991 | 0.9438 |
| 0.0068 | 544.0 | 565216 | 0.4054 | 0.7862 | 0.8158 | 0.8007 | 0.9448 |
| 0.0066 | 545.0 | 566255 | 0.4296 | 0.7962 | 0.7985 | 0.7973 | 0.9425 |
| 0.0064 | 546.0 | 567294 | 0.3988 | 0.7975 | 0.8158 | 0.8066 | 0.9449 |
| 0.0067 | 547.0 | 568333 | 0.4084 | 0.7894 | 0.8237 | 0.8062 | 0.9458 |
| 0.0057 | 548.0 | 569372 | 0.4053 | 0.7879 | 0.8203 | 0.8038 | 0.9453 |
| 0.0055 | 549.0 | 570411 | 0.3914 | 0.7887 | 0.8107 | 0.7996 | 0.9450 |
| 0.0066 | 550.0 | 571450 | 0.4108 | 0.7782 | 0.8212 | 0.7991 | 0.9447 |
| 0.0057 | 551.0 | 572489 | 0.4022 | 0.7943 | 0.8071 | 0.8006 | 0.9440 |
| 0.0063 | 552.0 | 573528 | 0.3876 | 0.7783 | 0.8233 | 0.8002 | 0.9451 |
| 0.0062 | 553.0 | 574567 | 0.3941 | 0.7911 | 0.8066 | 0.7988 | 0.9442 |
| 0.0059 | 554.0 | 575606 | 0.3945 | 0.7741 | 0.8158 | 0.7944 | 0.9441 |
| 0.0053 | 555.0 | 576645 | 0.4183 | 0.7928 | 0.7943 | 0.7936 | 0.9430 |
| 0.0062 | 556.0 | 577684 | 0.4055 | 0.7847 | 0.8152 | 0.7997 | 0.9446 |
| 0.0057 | 557.0 | 578723 | 0.3932 | 0.7923 | 0.8044 | 0.7983 | 0.9443 |
| 0.0065 | 558.0 | 579762 | 0.4124 | 0.7929 | 0.8046 | 0.7987 | 0.9438 |
| 0.0063 | 559.0 | 580801 | 0.4255 | 0.7847 | 0.8106 | 0.7974 | 0.9445 |
| 0.0066 | 560.0 | 581840 | 0.3730 | 0.7931 | 0.8151 | 0.8039 | 0.9458 |
| 0.0062 | 561.0 | 582879 | 0.4098 | 0.7852 | 0.8134 | 0.7990 | 0.9448 |
| 0.0056 | 562.0 | 583918 | 0.3879 | 0.7920 | 0.8059 | 0.7989 | 0.9447 |
| 0.0051 | 563.0 | 584957 | 0.4032 | 0.7792 | 0.8199 | 0.7991 | 0.9440 |
| 0.0059 | 564.0 | 585996 | 0.3985 | 0.7922 | 0.8075 | 0.7998 | 0.9445 |
| 0.006 | 565.0 | 587035 | 0.4118 | 0.7936 | 0.8041 | 0.7988 | 0.9442 |
| 0.0068 | 566.0 | 588074 | 0.4133 | 0.7855 | 0.8237 | 0.8041 | 0.9455 |
| 0.0058 | 567.0 | 589113 | 0.4040 | 0.7901 | 0.8066 | 0.7983 | 0.9441 |
| 0.006 | 568.0 | 590152 | 0.3931 | 0.7941 | 0.8082 | 0.8011 | 0.9447 |
| 0.0056 | 569.0 | 591191 | 0.4148 | 0.7845 | 0.7907 | 0.7876 | 0.9406 |
| 0.0057 | 570.0 | 592230 | 0.3904 | 0.7851 | 0.8080 | 0.7964 | 0.9443 |
| 0.007 | 571.0 | 593269 | 0.3987 | 0.7922 | 0.8032 | 0.7976 | 0.9442 |
| 0.007 | 572.0 | 594308 | 0.3963 | 0.7848 | 0.8194 | 0.8017 | 0.9461 |
| 0.0051 | 573.0 | 595347 | 0.4039 | 0.7777 | 0.8064 | 0.7918 | 0.9436 |
| 0.0052 | 574.0 | 596386 | 0.4090 | 0.7931 | 0.7974 | 0.7953 | 0.9427 |
| 0.005 | 575.0 | 597425 | 0.4270 | 0.7770 | 0.8100 | 0.7932 | 0.9430 |
| 0.006 | 576.0 | 598464 | 0.4374 | 0.7952 | 0.8070 | 0.8010 | 0.9442 |
| 0.0069 | 577.0 | 599503 | 0.3872 | 0.7906 | 0.8120 | 0.8012 | 0.9449 |
| 0.0067 | 578.0 | 600542 | 0.4086 | 0.7919 | 0.8129 | 0.8023 | 0.9453 |
| 0.005 | 579.0 | 601581 | 0.4135 | 0.7948 | 0.8160 | 0.8052 | 0.9455 |
| 0.0058 | 580.0 | 602620 | 0.4064 | 0.7973 | 0.7938 | 0.7955 | 0.9430 |
| 0.0059 | 581.0 | 603659 | 0.4099 | 0.7924 | 0.8194 | 0.8057 | 0.9461 |
| 0.0055 | 582.0 | 604698 | 0.4078 | 0.7927 | 0.8172 | 0.8048 | 0.9464 |
| 0.0053 | 583.0 | 605737 | 0.4350 | 0.7821 | 0.8053 | 0.7935 | 0.9432 |
| 0.0053 | 584.0 | 606776 | 0.3975 | 0.7950 | 0.8116 | 0.8032 | 0.9452 |
| 0.0048 | 585.0 | 607815 | 0.4141 | 0.7831 | 0.8160 | 0.7992 | 0.9440 |
| 0.0051 | 586.0 | 608854 | 0.3891 | 0.7865 | 0.8185 | 0.8022 | 0.9457 |
| 0.0043 | 587.0 | 609893 | 0.3943 | 0.7892 | 0.8086 | 0.7988 | 0.9443 |
| 0.0058 | 588.0 | 610932 | 0.4133 | 0.7885 | 0.8017 | 0.7951 | 0.9440 |
| 0.0055 | 589.0 | 611971 | 0.3794 | 0.7884 | 0.8012 | 0.7947 | 0.9430 |
| 0.0061 | 590.0 | 613010 | 0.4011 | 0.7903 | 0.8131 | 0.8015 | 0.9447 |
| 0.0046 | 591.0 | 614049 | 0.4261 | 0.7924 | 0.8151 | 0.8036 | 0.9444 |
| 0.0058 | 592.0 | 615088 | 0.4138 | 0.7895 | 0.8071 | 0.7982 | 0.9438 |
| 0.0057 | 593.0 | 616127 | 0.4321 | 0.7894 | 0.8041 | 0.7967 | 0.9427 |
| 0.0046 | 594.0 | 617166 | 0.4165 | 0.7903 | 0.8037 | 0.7970 | 0.9440 |
| 0.0059 | 595.0 | 618205 | 0.4174 | 0.7877 | 0.8147 | 0.8010 | 0.9454 |
| 0.004 | 596.0 | 619244 | 0.4101 | 0.7920 | 0.8188 | 0.8052 | 0.9464 |
| 0.0061 | 597.0 | 620283 | 0.4158 | 0.7775 | 0.8230 | 0.7996 | 0.9451 |
| 0.0052 | 598.0 | 621322 | 0.4135 | 0.7916 | 0.8079 | 0.7996 | 0.9440 |
| 0.0053 | 599.0 | 622361 | 0.4278 | 0.7786 | 0.8158 | 0.7968 | 0.9434 |
| 0.0047 | 600.0 | 623400 | 0.4131 | 0.7932 | 0.8205 | 0.8066 | 0.9456 |
| 0.0049 | 601.0 | 624439 | 0.4169 | 0.7903 | 0.8122 | 0.8011 | 0.9448 |
| 0.0055 | 602.0 | 625478 | 0.4176 | 0.7872 | 0.8203 | 0.8034 | 0.9457 |
| 0.0048 | 603.0 | 626517 | 0.4157 | 0.7869 | 0.8165 | 0.8014 | 0.9458 |
| 0.0061 | 604.0 | 627556 | 0.4003 | 0.7920 | 0.8142 | 0.8029 | 0.9456 |
| 0.0046 | 605.0 | 628595 | 0.4210 | 0.7742 | 0.8109 | 0.7922 | 0.9430 |
| 0.005 | 606.0 | 629634 | 0.3944 | 0.7856 | 0.8082 | 0.7967 | 0.9441 |
| 0.0047 | 607.0 | 630673 | 0.4046 | 0.7858 | 0.8138 | 0.7995 | 0.9447 |
| 0.0048 | 608.0 | 631712 | 0.4203 | 0.7907 | 0.8033 | 0.7970 | 0.9446 |
| 0.0051 | 609.0 | 632751 | 0.4583 | 0.7886 | 0.8151 | 0.8016 | 0.9445 |
| 0.0056 | 610.0 | 633790 | 0.4231 | 0.7867 | 0.8098 | 0.7981 | 0.9441 |
| 0.0044 | 611.0 | 634829 | 0.4289 | 0.7819 | 0.8019 | 0.7918 | 0.9436 |
| 0.0053 | 612.0 | 635868 | 0.4020 | 0.7925 | 0.8082 | 0.8003 | 0.9440 |
| 0.0052 | 613.0 | 636907 | 0.3960 | 0.7872 | 0.8140 | 0.8004 | 0.9449 |
| 0.0047 | 614.0 | 637946 | 0.4419 | 0.7830 | 0.8064 | 0.7945 | 0.9442 |
| 0.0052 | 615.0 | 638985 | 0.4324 | 0.7875 | 0.8235 | 0.8051 | 0.9454 |
| 0.0047 | 616.0 | 640024 | 0.4212 | 0.7976 | 0.8111 | 0.8043 | 0.9452 |
| 0.005 | 617.0 | 641063 | 0.4215 | 0.7875 | 0.8203 | 0.8036 | 0.9452 |
| 0.0042 | 618.0 | 642102 | 0.4293 | 0.7951 | 0.8084 | 0.8017 | 0.9443 |
| 0.0047 | 619.0 | 643141 | 0.4402 | 0.7930 | 0.8161 | 0.8044 | 0.9461 |
| 0.0051 | 620.0 | 644180 | 0.4092 | 0.7921 | 0.8080 | 0.8000 | 0.9448 |
| 0.005 | 621.0 | 645219 | 0.4187 | 0.7927 | 0.8174 | 0.8049 | 0.9455 |
| 0.0048 | 622.0 | 646258 | 0.4018 | 0.7919 | 0.8169 | 0.8042 | 0.9461 |
| 0.0049 | 623.0 | 647297 | 0.4088 | 0.7915 | 0.8235 | 0.8072 | 0.9465 |
| 0.0054 | 624.0 | 648336 | 0.4178 | 0.7859 | 0.8152 | 0.8003 | 0.9453 |
| 0.0038 | 625.0 | 649375 | 0.4053 | 0.7976 | 0.8041 | 0.8008 | 0.9451 |
| 0.0052 | 626.0 | 650414 | 0.4076 | 0.7961 | 0.8091 | 0.8025 | 0.9449 |
| 0.0037 | 627.0 | 651453 | 0.4179 | 0.7879 | 0.8102 | 0.7989 | 0.9430 |
| 0.004 | 628.0 | 652492 | 0.4280 | 0.7844 | 0.8061 | 0.7951 | 0.9439 |
| 0.0039 | 629.0 | 653531 | 0.4373 | 0.7895 | 0.8215 | 0.8052 | 0.9467 |
| 0.005 | 630.0 | 654570 | 0.4213 | 0.7868 | 0.8116 | 0.7990 | 0.9453 |
| 0.0049 | 631.0 | 655609 | 0.4155 | 0.7908 | 0.8082 | 0.7994 | 0.9460 |
| 0.0043 | 632.0 | 656648 | 0.4265 | 0.7962 | 0.8086 | 0.8024 | 0.9449 |
| 0.0037 | 633.0 | 657687 | 0.4460 | 0.7902 | 0.8098 | 0.7999 | 0.9457 |
| 0.0049 | 634.0 | 658726 | 0.4411 | 0.7893 | 0.8082 | 0.7986 | 0.9451 |
| 0.0042 | 635.0 | 659765 | 0.4256 | 0.7845 | 0.8091 | 0.7966 | 0.9447 |
| 0.0042 | 636.0 | 660804 | 0.4087 | 0.7842 | 0.8169 | 0.8002 | 0.9453 |
| 0.005 | 637.0 | 661843 | 0.4191 | 0.7933 | 0.8140 | 0.8035 | 0.9463 |
| 0.0036 | 638.0 | 662882 | 0.4556 | 0.7807 | 0.8221 | 0.8009 | 0.9450 |
| 0.0041 | 639.0 | 663921 | 0.4145 | 0.7937 | 0.8111 | 0.8023 | 0.9454 |
| 0.004 | 640.0 | 664960 | 0.4288 | 0.7848 | 0.8093 | 0.7969 | 0.9445 |
| 0.0039 | 641.0 | 665999 | 0.4017 | 0.7897 | 0.8113 | 0.8003 | 0.9461 |
| 0.0044 | 642.0 | 667038 | 0.4181 | 0.7884 | 0.8084 | 0.7983 | 0.9451 |
| 0.0042 | 643.0 | 668077 | 0.4261 | 0.7896 | 0.8124 | 0.8008 | 0.9446 |
| 0.004 | 644.0 | 669116 | 0.4156 | 0.7939 | 0.8208 | 0.8072 | 0.9469 |
| 0.0042 | 645.0 | 670155 | 0.4201 | 0.7892 | 0.8232 | 0.8058 | 0.9460 |
| 0.0044 | 646.0 | 671194 | 0.4345 | 0.7943 | 0.8124 | 0.8032 | 0.9448 |
| 0.004 | 647.0 | 672233 | 0.4265 | 0.7944 | 0.8143 | 0.8043 | 0.9460 |
| 0.0036 | 648.0 | 673272 | 0.4316 | 0.7810 | 0.8304 | 0.8049 | 0.9462 |
| 0.0039 | 649.0 | 674311 | 0.4066 | 0.7922 | 0.8115 | 0.8017 | 0.9452 |
| 0.0048 | 650.0 | 675350 | 0.4306 | 0.7909 | 0.8223 | 0.8063 | 0.9458 |
| 0.0039 | 651.0 | 676389 | 0.4231 | 0.8006 | 0.8179 | 0.8092 | 0.9459 |
| 0.0042 | 652.0 | 677428 | 0.4020 | 0.7900 | 0.8169 | 0.8032 | 0.9452 |
| 0.0031 | 653.0 | 678467 | 0.3917 | 0.7954 | 0.8161 | 0.8056 | 0.9462 |
| 0.0041 | 654.0 | 679506 | 0.4286 | 0.7932 | 0.8156 | 0.8042 | 0.9461 |
| 0.0049 | 655.0 | 680545 | 0.3924 | 0.7934 | 0.8161 | 0.8046 | 0.9453 |
| 0.003 | 656.0 | 681584 | 0.4151 | 0.7940 | 0.8163 | 0.8050 | 0.9461 |
| 0.0037 | 657.0 | 682623 | 0.4409 | 0.7925 | 0.8192 | 0.8056 | 0.9457 |
| 0.0041 | 658.0 | 683662 | 0.4226 | 0.7951 | 0.8104 | 0.8026 | 0.9447 |
| 0.004 | 659.0 | 684701 | 0.4222 | 0.7978 | 0.8102 | 0.8040 | 0.9457 |
| 0.004 | 660.0 | 685740 | 0.4003 | 0.7958 | 0.8178 | 0.8066 | 0.9473 |
| 0.0038 | 661.0 | 686779 | 0.4494 | 0.7893 | 0.8088 | 0.7989 | 0.9445 |
| 0.0043 | 662.0 | 687818 | 0.4267 | 0.7914 | 0.8197 | 0.8053 | 0.9466 |
| 0.0032 | 663.0 | 688857 | 0.4261 | 0.7871 | 0.8167 | 0.8016 | 0.9460 |
| 0.0044 | 664.0 | 689896 | 0.4473 | 0.7904 | 0.8178 | 0.8039 | 0.9460 |
| 0.0035 | 665.0 | 690935 | 0.4303 | 0.7942 | 0.8145 | 0.8042 | 0.9450 |
| 0.0039 | 666.0 | 691974 | 0.4298 | 0.8003 | 0.8075 | 0.8039 | 0.9454 |
| 0.0034 | 667.0 | 693013 | 0.4435 | 0.7843 | 0.8203 | 0.8019 | 0.9458 |
| 0.0035 | 668.0 | 694052 | 0.4583 | 0.7872 | 0.8160 | 0.8013 | 0.9450 |
| 0.0038 | 669.0 | 695091 | 0.4348 | 0.7885 | 0.8098 | 0.7990 | 0.9446 |
| 0.0041 | 670.0 | 696130 | 0.4066 | 0.7887 | 0.8304 | 0.8090 | 0.9472 |
| 0.0038 | 671.0 | 697169 | 0.4265 | 0.7963 | 0.8232 | 0.8095 | 0.9476 |
| 0.0039 | 672.0 | 698208 | 0.4267 | 0.7927 | 0.8158 | 0.8040 | 0.9451 |
| 0.0039 | 673.0 | 699247 | 0.4245 | 0.7972 | 0.8185 | 0.8077 | 0.9468 |
| 0.0042 | 674.0 | 700286 | 0.4148 | 0.7990 | 0.8183 | 0.8085 | 0.9468 |
| 0.0032 | 675.0 | 701325 | 0.4403 | 0.7924 | 0.8124 | 0.8022 | 0.9458 |
| 0.0036 | 676.0 | 702364 | 0.4177 | 0.7879 | 0.8230 | 0.8051 | 0.9470 |
| 0.0033 | 677.0 | 703403 | 0.4098 | 0.7990 | 0.8277 | 0.8131 | 0.9473 |
| 0.0028 | 678.0 | 704442 | 0.4441 | 0.7920 | 0.8188 | 0.8052 | 0.9462 |
| 0.0034 | 679.0 | 705481 | 0.4288 | 0.7942 | 0.8167 | 0.8053 | 0.9461 |
| 0.0032 | 680.0 | 706520 | 0.4593 | 0.7868 | 0.8188 | 0.8025 | 0.9454 |
| 0.0035 | 681.0 | 707559 | 0.4404 | 0.7984 | 0.8116 | 0.8050 | 0.9456 |
| 0.0037 | 682.0 | 708598 | 0.4237 | 0.7929 | 0.8203 | 0.8063 | 0.9470 |
| 0.0035 | 683.0 | 709637 | 0.4301 | 0.7901 | 0.8217 | 0.8056 | 0.9464 |
| 0.004 | 684.0 | 710676 | 0.4287 | 0.7920 | 0.8181 | 0.8049 | 0.9465 |
| 0.0031 | 685.0 | 711715 | 0.4348 | 0.7886 | 0.8163 | 0.8022 | 0.9458 |
| 0.0031 | 686.0 | 712754 | 0.4507 | 0.7810 | 0.8201 | 0.8001 | 0.9453 |
| 0.0038 | 687.0 | 713793 | 0.4299 | 0.8003 | 0.8140 | 0.8071 | 0.9463 |
| 0.0041 | 688.0 | 714832 | 0.4361 | 0.7915 | 0.8210 | 0.8060 | 0.9466 |
| 0.0039 | 689.0 | 715871 | 0.4413 | 0.7987 | 0.8122 | 0.8054 | 0.9447 |
| 0.0027 | 690.0 | 716910 | 0.4446 | 0.7958 | 0.8136 | 0.8046 | 0.9453 |
| 0.0031 | 691.0 | 717949 | 0.4536 | 0.7978 | 0.8151 | 0.8063 | 0.9464 |
| 0.0036 | 692.0 | 718988 | 0.4469 | 0.7924 | 0.8133 | 0.8027 | 0.9452 |
| 0.0031 | 693.0 | 720027 | 0.4605 | 0.7945 | 0.8106 | 0.8025 | 0.9453 |
| 0.003 | 694.0 | 721066 | 0.4526 | 0.7932 | 0.8111 | 0.8021 | 0.9441 |
| 0.0036 | 695.0 | 722105 | 0.4492 | 0.7926 | 0.8208 | 0.8064 | 0.9467 |
| 0.003 | 696.0 | 723144 | 0.4359 | 0.7911 | 0.8206 | 0.8056 | 0.9463 |
| 0.0041 | 697.0 | 724183 | 0.4336 | 0.7948 | 0.8210 | 0.8077 | 0.9459 |
| 0.003 | 698.0 | 725222 | 0.4505 | 0.7930 | 0.8115 | 0.8021 | 0.9454 |
| 0.0034 | 699.0 | 726261 | 0.4347 | 0.7910 | 0.8106 | 0.8007 | 0.9448 |
| 0.0025 | 700.0 | 727300 | 0.4275 | 0.7909 | 0.8253 | 0.8077 | 0.9462 |
| 0.003 | 701.0 | 728339 | 0.4469 | 0.7893 | 0.8244 | 0.8065 | 0.9461 |
| 0.0038 | 702.0 | 729378 | 0.4315 | 0.7928 | 0.8109 | 0.8017 | 0.9440 |
| 0.0036 | 703.0 | 730417 | 0.4584 | 0.7873 | 0.8187 | 0.8027 | 0.9454 |
| 0.0028 | 704.0 | 731456 | 0.4466 | 0.7980 | 0.8082 | 0.8031 | 0.9448 |
| 0.0033 | 705.0 | 732495 | 0.4314 | 0.7799 | 0.8102 | 0.7947 | 0.9445 |
| 0.0035 | 706.0 | 733534 | 0.4332 | 0.7943 | 0.8048 | 0.7995 | 0.9448 |
| 0.0031 | 707.0 | 734573 | 0.4343 | 0.7937 | 0.8118 | 0.8026 | 0.9450 |
| 0.0032 | 708.0 | 735612 | 0.4278 | 0.7934 | 0.8187 | 0.8058 | 0.9458 |
| 0.0027 | 709.0 | 736651 | 0.4521 | 0.7909 | 0.8201 | 0.8052 | 0.9462 |
| 0.003 | 710.0 | 737690 | 0.4406 | 0.7853 | 0.8192 | 0.8019 | 0.9451 |
| 0.0038 | 711.0 | 738729 | 0.4325 | 0.7882 | 0.8224 | 0.8050 | 0.9450 |
| 0.0022 | 712.0 | 739768 | 0.4471 | 0.7984 | 0.8199 | 0.8090 | 0.9458 |
| 0.0037 | 713.0 | 740807 | 0.4173 | 0.7936 | 0.8131 | 0.8032 | 0.9445 |
| 0.003 | 714.0 | 741846 | 0.4355 | 0.7904 | 0.8201 | 0.8049 | 0.9454 |
| 0.0032 | 715.0 | 742885 | 0.4345 | 0.7967 | 0.8167 | 0.8066 | 0.9458 |
| 0.0031 | 716.0 | 743924 | 0.4513 | 0.7849 | 0.8131 | 0.7988 | 0.9448 |
| 0.0028 | 717.0 | 744963 | 0.4552 | 0.7731 | 0.8062 | 0.7893 | 0.9429 |
| 0.0033 | 718.0 | 746002 | 0.4608 | 0.7810 | 0.8077 | 0.7941 | 0.9434 |
| 0.0026 | 719.0 | 747041 | 0.4363 | 0.7904 | 0.8223 | 0.8060 | 0.9464 |
| 0.0024 | 720.0 | 748080 | 0.4570 | 0.7941 | 0.8235 | 0.8085 | 0.9453 |
| 0.0028 | 721.0 | 749119 | 0.4339 | 0.7926 | 0.8185 | 0.8054 | 0.9459 |
| 0.0032 | 722.0 | 750158 | 0.4595 | 0.7942 | 0.8188 | 0.8063 | 0.9455 |
| 0.003 | 723.0 | 751197 | 0.4534 | 0.7906 | 0.8165 | 0.8033 | 0.9455 |
| 0.003 | 724.0 | 752236 | 0.4430 | 0.7926 | 0.8125 | 0.8024 | 0.9444 |
| 0.0029 | 725.0 | 753275 | 0.4267 | 0.7960 | 0.8208 | 0.8082 | 0.9460 |
| 0.0022 | 726.0 | 754314 | 0.4292 | 0.7913 | 0.8196 | 0.8052 | 0.9458 |
| 0.0032 | 727.0 | 755353 | 0.4656 | 0.7909 | 0.8160 | 0.8032 | 0.9446 |
| 0.0029 | 728.0 | 756392 | 0.4333 | 0.7932 | 0.8129 | 0.8029 | 0.9453 |
| 0.0027 | 729.0 | 757431 | 0.4440 | 0.7939 | 0.8197 | 0.8066 | 0.9461 |
| 0.0027 | 730.0 | 758470 | 0.4384 | 0.7948 | 0.8219 | 0.8081 | 0.9458 |
| 0.0024 | 731.0 | 759509 | 0.4551 | 0.7966 | 0.8196 | 0.8079 | 0.9467 |
| 0.0022 | 732.0 | 760548 | 0.4627 | 0.7985 | 0.8219 | 0.8100 | 0.9471 |
| 0.0028 | 733.0 | 761587 | 0.4318 | 0.8007 | 0.8190 | 0.8098 | 0.9468 |
| 0.0029 | 734.0 | 762626 | 0.4360 | 0.8006 | 0.8208 | 0.8106 | 0.9464 |
| 0.0028 | 735.0 | 763665 | 0.4450 | 0.7955 | 0.8118 | 0.8036 | 0.9455 |
| 0.0025 | 736.0 | 764704 | 0.4452 | 0.7965 | 0.8232 | 0.8096 | 0.9460 |
| 0.0024 | 737.0 | 765743 | 0.4451 | 0.7930 | 0.8205 | 0.8065 | 0.9463 |
| 0.0024 | 738.0 | 766782 | 0.4868 | 0.7818 | 0.8286 | 0.8045 | 0.9453 |
| 0.0022 | 739.0 | 767821 | 0.4549 | 0.7953 | 0.8223 | 0.8086 | 0.9459 |
| 0.0024 | 740.0 | 768860 | 0.4878 | 0.7934 | 0.8097 | 0.8014 | 0.9451 |
| 0.0024 | 741.0 | 769899 | 0.4503 | 0.7944 | 0.8185 | 0.8063 | 0.9457 |
| 0.0032 | 742.0 | 770938 | 0.4700 | 0.7957 | 0.8185 | 0.8069 | 0.9454 |
| 0.003 | 743.0 | 771977 | 0.4781 | 0.7889 | 0.8163 | 0.8024 | 0.9453 |
| 0.0027 | 744.0 | 773016 | 0.4661 | 0.7956 | 0.8217 | 0.8085 | 0.9463 |
| 0.003 | 745.0 | 774055 | 0.4623 | 0.7993 | 0.8269 | 0.8129 | 0.9472 |
| 0.0032 | 746.0 | 775094 | 0.4405 | 0.8011 | 0.8230 | 0.8119 | 0.9470 |
| 0.0024 | 747.0 | 776133 | 0.4544 | 0.7884 | 0.8199 | 0.8038 | 0.9456 |
| 0.0027 | 748.0 | 777172 | 0.4450 | 0.7967 | 0.8152 | 0.8059 | 0.9463 |
| 0.0025 | 749.0 | 778211 | 0.4320 | 0.7978 | 0.8197 | 0.8086 | 0.9471 |
| 0.0026 | 750.0 | 779250 | 0.4498 | 0.7963 | 0.8262 | 0.8110 | 0.9462 |
| 0.0022 | 751.0 | 780289 | 0.4660 | 0.7972 | 0.8190 | 0.8080 | 0.9460 |
| 0.0028 | 752.0 | 781328 | 0.4457 | 0.7964 | 0.8192 | 0.8076 | 0.9466 |
| 0.0026 | 753.0 | 782367 | 0.4408 | 0.7994 | 0.8275 | 0.8132 | 0.9470 |
| 0.0027 | 754.0 | 783406 | 0.4623 | 0.7953 | 0.8075 | 0.8014 | 0.9449 |
| 0.002 | 755.0 | 784445 | 0.4495 | 0.7948 | 0.8156 | 0.8051 | 0.9461 |
| 0.0025 | 756.0 | 785484 | 0.4674 | 0.7939 | 0.8205 | 0.8069 | 0.9454 |
| 0.0023 | 757.0 | 786523 | 0.4515 | 0.7963 | 0.8077 | 0.8020 | 0.9456 |
| 0.0028 | 758.0 | 787562 | 0.4399 | 0.7977 | 0.8206 | 0.8090 | 0.9471 |
| 0.0024 | 759.0 | 788601 | 0.4261 | 0.7936 | 0.8149 | 0.8041 | 0.9456 |
| 0.002 | 760.0 | 789640 | 0.4786 | 0.7961 | 0.8107 | 0.8034 | 0.9456 |
| 0.0027 | 761.0 | 790679 | 0.4673 | 0.7930 | 0.8134 | 0.8031 | 0.9454 |
| 0.0021 | 762.0 | 791718 | 0.4832 | 0.7869 | 0.8145 | 0.8005 | 0.9452 |
| 0.0018 | 763.0 | 792757 | 0.4918 | 0.7971 | 0.8221 | 0.8094 | 0.9472 |
| 0.0018 | 764.0 | 793796 | 0.5213 | 0.7918 | 0.8102 | 0.8009 | 0.9448 |
| 0.0022 | 765.0 | 794835 | 0.4741 | 0.7951 | 0.8113 | 0.8031 | 0.9457 |
| 0.0021 | 766.0 | 795874 | 0.4708 | 0.7916 | 0.8201 | 0.8056 | 0.9463 |
| 0.0035 | 767.0 | 796913 | 0.4740 | 0.7942 | 0.8181 | 0.8060 | 0.9464 |
| 0.0024 | 768.0 | 797952 | 0.4632 | 0.7915 | 0.8109 | 0.8011 | 0.9457 |
| 0.0019 | 769.0 | 798991 | 0.4757 | 0.7960 | 0.8116 | 0.8037 | 0.9466 |
| 0.0023 | 770.0 | 800030 | 0.4700 | 0.7860 | 0.8196 | 0.8024 | 0.9468 |
| 0.0013 | 771.0 | 801069 | 0.4949 | 0.7938 | 0.8242 | 0.8087 | 0.9478 |
| 0.002 | 772.0 | 802108 | 0.4691 | 0.7976 | 0.8196 | 0.8084 | 0.9468 |
| 0.0022 | 773.0 | 803147 | 0.4645 | 0.7860 | 0.8161 | 0.8008 | 0.9459 |
| 0.0023 | 774.0 | 804186 | 0.4572 | 0.7911 | 0.8129 | 0.8018 | 0.9462 |
| 0.0021 | 775.0 | 805225 | 0.4681 | 0.8004 | 0.8042 | 0.8023 | 0.9454 |
| 0.0018 | 776.0 | 806264 | 0.4749 | 0.7894 | 0.8241 | 0.8063 | 0.9464 |
| 0.0018 | 777.0 | 807303 | 0.4586 | 0.7931 | 0.8145 | 0.8037 | 0.9460 |
| 0.0026 | 778.0 | 808342 | 0.4502 | 0.7997 | 0.8088 | 0.8042 | 0.9457 |
| 0.0022 | 779.0 | 809381 | 0.4517 | 0.7887 | 0.8169 | 0.8025 | 0.9450 |
| 0.0018 | 780.0 | 810420 | 0.4538 | 0.7916 | 0.8118 | 0.8016 | 0.9452 |
| 0.0017 | 781.0 | 811459 | 0.4789 | 0.7941 | 0.8097 | 0.8018 | 0.9448 |
| 0.0017 | 782.0 | 812498 | 0.4762 | 0.7916 | 0.8138 | 0.8025 | 0.9454 |
| 0.0019 | 783.0 | 813537 | 0.4563 | 0.7959 | 0.8230 | 0.8092 | 0.9466 |
| 0.0025 | 784.0 | 814576 | 0.4696 | 0.792 | 0.8201 | 0.8058 | 0.9452 |
| 0.0017 | 785.0 | 815615 | 0.4835 | 0.7997 | 0.8152 | 0.8074 | 0.9459 |
| 0.0016 | 786.0 | 816654 | 0.4955 | 0.7944 | 0.8160 | 0.8050 | 0.9454 |
| 0.0021 | 787.0 | 817693 | 0.4741 | 0.7928 | 0.8140 | 0.8033 | 0.9446 |
| 0.0015 | 788.0 | 818732 | 0.4985 | 0.7845 | 0.8136 | 0.7988 | 0.9443 |
| 0.0023 | 789.0 | 819771 | 0.4671 | 0.7954 | 0.8079 | 0.8016 | 0.9452 |
| 0.0024 | 790.0 | 820810 | 0.4795 | 0.7950 | 0.8095 | 0.8022 | 0.9453 |
| 0.0018 | 791.0 | 821849 | 0.4902 | 0.7925 | 0.8145 | 0.8034 | 0.9453 |
| 0.0017 | 792.0 | 822888 | 0.4746 | 0.7907 | 0.8125 | 0.8015 | 0.9454 |
| 0.0018 | 793.0 | 823927 | 0.4977 | 0.7812 | 0.8104 | 0.7955 | 0.9431 |
| 0.0022 | 794.0 | 824966 | 0.4832 | 0.7925 | 0.8062 | 0.7993 | 0.9439 |
| 0.0022 | 795.0 | 826005 | 0.4734 | 0.7775 | 0.8053 | 0.7912 | 0.9424 |
| 0.0018 | 796.0 | 827044 | 0.4914 | 0.7968 | 0.8093 | 0.8030 | 0.9449 |
| 0.0016 | 797.0 | 828083 | 0.4924 | 0.7886 | 0.8109 | 0.7996 | 0.9446 |
| 0.0015 | 798.0 | 829122 | 0.5124 | 0.7819 | 0.8232 | 0.8020 | 0.9452 |
| 0.0018 | 799.0 | 830161 | 0.4870 | 0.7924 | 0.8226 | 0.8072 | 0.9465 |
| 0.0013 | 800.0 | 831200 | 0.4897 | 0.7949 | 0.8152 | 0.8049 | 0.9459 |
| 0.002 | 801.0 | 832239 | 0.4733 | 0.7920 | 0.8199 | 0.8057 | 0.9465 |
| 0.0019 | 802.0 | 833278 | 0.4816 | 0.8005 | 0.8109 | 0.8057 | 0.9461 |
| 0.0016 | 803.0 | 834317 | 0.5166 | 0.7931 | 0.8071 | 0.8001 | 0.9441 |
| 0.0023 | 804.0 | 835356 | 0.4659 | 0.8043 | 0.8154 | 0.8098 | 0.9472 |
| 0.0017 | 805.0 | 836395 | 0.5199 | 0.7960 | 0.8163 | 0.8060 | 0.9460 |
| 0.0019 | 806.0 | 837434 | 0.5119 | 0.8 | 0.8269 | 0.8132 | 0.9470 |
| 0.0015 | 807.0 | 838473 | 0.4911 | 0.7910 | 0.8174 | 0.8040 | 0.9452 |
| 0.0015 | 808.0 | 839512 | 0.5130 | 0.7922 | 0.8196 | 0.8056 | 0.9459 |
| 0.0021 | 809.0 | 840551 | 0.4682 | 0.7976 | 0.8239 | 0.8105 | 0.9467 |
| 0.0012 | 810.0 | 841590 | 0.4880 | 0.7930 | 0.8181 | 0.8054 | 0.9458 |
| 0.0023 | 811.0 | 842629 | 0.4703 | 0.7934 | 0.8172 | 0.8051 | 0.9455 |
| 0.0016 | 812.0 | 843668 | 0.4727 | 0.7925 | 0.8219 | 0.8069 | 0.9468 |
| 0.0017 | 813.0 | 844707 | 0.4682 | 0.7951 | 0.8185 | 0.8066 | 0.9464 |
| 0.0014 | 814.0 | 845746 | 0.4761 | 0.7973 | 0.8120 | 0.8046 | 0.9451 |
| 0.0014 | 815.0 | 846785 | 0.4847 | 0.7929 | 0.8205 | 0.8064 | 0.9456 |
| 0.0013 | 816.0 | 847824 | 0.4930 | 0.7869 | 0.8091 | 0.7978 | 0.9442 |
| 0.0017 | 817.0 | 848863 | 0.4857 | 0.7924 | 0.8097 | 0.8009 | 0.9445 |
| 0.0014 | 818.0 | 849902 | 0.5136 | 0.7914 | 0.8098 | 0.8005 | 0.9440 |
| 0.0017 | 819.0 | 850941 | 0.5051 | 0.7877 | 0.8044 | 0.7960 | 0.9438 |
| 0.0013 | 820.0 | 851980 | 0.4995 | 0.7983 | 0.8089 | 0.8036 | 0.9438 |
| 0.0017 | 821.0 | 853019 | 0.4958 | 0.7937 | 0.8170 | 0.8052 | 0.9446 |
| 0.0016 | 822.0 | 854058 | 0.4868 | 0.7949 | 0.8167 | 0.8056 | 0.9458 |
| 0.0018 | 823.0 | 855097 | 0.4700 | 0.7900 | 0.8203 | 0.8048 | 0.9453 |
| 0.0013 | 824.0 | 856136 | 0.4725 | 0.7972 | 0.8233 | 0.8101 | 0.9470 |
| 0.0018 | 825.0 | 857175 | 0.5226 | 0.8029 | 0.8167 | 0.8097 | 0.9462 |
| 0.0016 | 826.0 | 858214 | 0.4596 | 0.7897 | 0.8127 | 0.8010 | 0.9452 |
| 0.0014 | 827.0 | 859253 | 0.4858 | 0.7960 | 0.8163 | 0.8060 | 0.9456 |
| 0.0013 | 828.0 | 860292 | 0.4978 | 0.7950 | 0.8257 | 0.8101 | 0.9461 |
| 0.0016 | 829.0 | 861331 | 0.4839 | 0.7964 | 0.8127 | 0.8045 | 0.9456 |
| 0.0009 | 830.0 | 862370 | 0.4902 | 0.7962 | 0.8174 | 0.8066 | 0.9456 |
| 0.0012 | 831.0 | 863409 | 0.4870 | 0.7976 | 0.8167 | 0.8070 | 0.9456 |
| 0.0016 | 832.0 | 864448 | 0.4764 | 0.8030 | 0.8154 | 0.8091 | 0.9455 |
| 0.001 | 833.0 | 865487 | 0.5190 | 0.7996 | 0.8140 | 0.8067 | 0.9446 |
| 0.0016 | 834.0 | 866526 | 0.4974 | 0.7908 | 0.8228 | 0.8065 | 0.9443 |
| 0.0013 | 835.0 | 867565 | 0.5002 | 0.7969 | 0.8190 | 0.8078 | 0.9461 |
| 0.0014 | 836.0 | 868604 | 0.4879 | 0.8002 | 0.8109 | 0.8055 | 0.9452 |
| 0.0014 | 837.0 | 869643 | 0.4731 | 0.7943 | 0.8131 | 0.8036 | 0.9448 |
| 0.0013 | 838.0 | 870682 | 0.4893 | 0.7951 | 0.8142 | 0.8045 | 0.9449 |
| 0.0012 | 839.0 | 871721 | 0.4717 | 0.7929 | 0.8082 | 0.8005 | 0.9449 |
| 0.0009 | 840.0 | 872760 | 0.5202 | 0.7976 | 0.8024 | 0.8000 | 0.9442 |
| 0.0014 | 841.0 | 873799 | 0.5112 | 0.7929 | 0.8188 | 0.8056 | 0.9455 |
| 0.0016 | 842.0 | 874838 | 0.4835 | 0.7985 | 0.8116 | 0.8050 | 0.9453 |
| 0.0012 | 843.0 | 875877 | 0.4781 | 0.8050 | 0.8244 | 0.8146 | 0.9471 |
| 0.0012 | 844.0 | 876916 | 0.4797 | 0.8014 | 0.8170 | 0.8092 | 0.9459 |
| 0.001 | 845.0 | 877955 | 0.4919 | 0.8047 | 0.8208 | 0.8127 | 0.9463 |
| 0.001 | 846.0 | 878994 | 0.4736 | 0.7971 | 0.8201 | 0.8085 | 0.9455 |
| 0.0013 | 847.0 | 880033 | 0.4985 | 0.7943 | 0.8190 | 0.8065 | 0.9456 |
| 0.0014 | 848.0 | 881072 | 0.5027 | 0.7992 | 0.8161 | 0.8076 | 0.9449 |
| 0.0012 | 849.0 | 882111 | 0.5396 | 0.7956 | 0.8091 | 0.8023 | 0.9445 |
| 0.0011 | 850.0 | 883150 | 0.5024 | 0.8031 | 0.8167 | 0.8098 | 0.9458 |
| 0.0015 | 851.0 | 884189 | 0.4753 | 0.8027 | 0.8174 | 0.8100 | 0.9459 |
| 0.001 | 852.0 | 885228 | 0.4900 | 0.8027 | 0.8169 | 0.8097 | 0.9454 |
| 0.001 | 853.0 | 886267 | 0.4898 | 0.7969 | 0.8201 | 0.8083 | 0.9461 |
| 0.0009 | 854.0 | 887306 | 0.5069 | 0.7961 | 0.8277 | 0.8116 | 0.9474 |
| 0.0011 | 855.0 | 888345 | 0.4986 | 0.8039 | 0.8181 | 0.8110 | 0.9467 |
| 0.0014 | 856.0 | 889384 | 0.4845 | 0.8009 | 0.8143 | 0.8076 | 0.9460 |
| 0.0011 | 857.0 | 890423 | 0.4704 | 0.7986 | 0.8239 | 0.8110 | 0.9470 |
| 0.0009 | 858.0 | 891462 | 0.4773 | 0.8015 | 0.8149 | 0.8081 | 0.9462 |
| 0.0009 | 859.0 | 892501 | 0.4992 | 0.8013 | 0.8230 | 0.8120 | 0.9470 |
| 0.0011 | 860.0 | 893540 | 0.5007 | 0.7990 | 0.8244 | 0.8115 | 0.9469 |
| 0.0012 | 861.0 | 894579 | 0.4777 | 0.7973 | 0.8251 | 0.8110 | 0.9472 |
| 0.0009 | 862.0 | 895618 | 0.5165 | 0.7955 | 0.8201 | 0.8076 | 0.9458 |
| 0.0011 | 863.0 | 896657 | 0.5115 | 0.7917 | 0.8235 | 0.8073 | 0.9462 |
| 0.0013 | 864.0 | 897696 | 0.5022 | 0.7986 | 0.8257 | 0.8119 | 0.9464 |
| 0.0009 | 865.0 | 898735 | 0.5226 | 0.7979 | 0.8287 | 0.8130 | 0.9468 |
| 0.0012 | 866.0 | 899774 | 0.4939 | 0.8002 | 0.8259 | 0.8128 | 0.9471 |
| 0.0009 | 867.0 | 900813 | 0.5016 | 0.8058 | 0.8248 | 0.8152 | 0.9476 |
| 0.001 | 868.0 | 901852 | 0.4974 | 0.8059 | 0.8232 | 0.8144 | 0.9471 |
| 0.0016 | 869.0 | 902891 | 0.4697 | 0.7994 | 0.8217 | 0.8104 | 0.9469 |
| 0.0011 | 870.0 | 903930 | 0.4896 | 0.7983 | 0.8239 | 0.8109 | 0.9470 |
| 0.0008 | 871.0 | 904969 | 0.4924 | 0.8019 | 0.8242 | 0.8129 | 0.9475 |
| 0.0009 | 872.0 | 906008 | 0.4997 | 0.8025 | 0.8181 | 0.8102 | 0.9468 |
| 0.0007 | 873.0 | 907047 | 0.5172 | 0.7912 | 0.8203 | 0.8055 | 0.9459 |
| 0.0009 | 874.0 | 908086 | 0.5148 | 0.7967 | 0.8194 | 0.8079 | 0.9462 |
| 0.0008 | 875.0 | 909125 | 0.5097 | 0.8034 | 0.8242 | 0.8137 | 0.9475 |
| 0.001 | 876.0 | 910164 | 0.4955 | 0.8002 | 0.8244 | 0.8121 | 0.9470 |
| 0.0012 | 877.0 | 911203 | 0.4828 | 0.8050 | 0.8239 | 0.8143 | 0.9468 |
| 0.001 | 878.0 | 912242 | 0.4906 | 0.7994 | 0.8152 | 0.8072 | 0.9464 |
| 0.0009 | 879.0 | 913281 | 0.5100 | 0.8028 | 0.8172 | 0.8099 | 0.9464 |
| 0.001 | 880.0 | 914320 | 0.4833 | 0.7968 | 0.8221 | 0.8093 | 0.9469 |
| 0.0009 | 881.0 | 915359 | 0.4925 | 0.7928 | 0.8147 | 0.8036 | 0.9455 |
| 0.0007 | 882.0 | 916398 | 0.5253 | 0.7965 | 0.8262 | 0.8111 | 0.9468 |
| 0.0006 | 883.0 | 917437 | 0.5365 | 0.7983 | 0.8253 | 0.8116 | 0.9465 |
| 0.0008 | 884.0 | 918476 | 0.5234 | 0.7961 | 0.8226 | 0.8091 | 0.9461 |
| 0.0007 | 885.0 | 919515 | 0.5323 | 0.8008 | 0.8161 | 0.8084 | 0.9452 |
| 0.0009 | 886.0 | 920554 | 0.5377 | 0.7877 | 0.8178 | 0.8024 | 0.9448 |
| 0.0008 | 887.0 | 921593 | 0.5289 | 0.7999 | 0.8215 | 0.8106 | 0.9466 |
| 0.0006 | 888.0 | 922632 | 0.5331 | 0.7976 | 0.8167 | 0.8070 | 0.9454 |
| 0.0007 | 889.0 | 923671 | 0.5433 | 0.8017 | 0.8134 | 0.8075 | 0.9459 |
| 0.0009 | 890.0 | 924710 | 0.5128 | 0.7992 | 0.8201 | 0.8095 | 0.9464 |
| 0.0008 | 891.0 | 925749 | 0.5092 | 0.8024 | 0.8212 | 0.8117 | 0.9457 |
| 0.0006 | 892.0 | 926788 | 0.5281 | 0.7970 | 0.8217 | 0.8092 | 0.9458 |
| 0.0009 | 893.0 | 927827 | 0.4930 | 0.8018 | 0.8187 | 0.8101 | 0.9463 |
| 0.0007 | 894.0 | 928866 | 0.5119 | 0.8013 | 0.8156 | 0.8084 | 0.9465 |
| 0.0008 | 895.0 | 929905 | 0.5281 | 0.7937 | 0.8275 | 0.8103 | 0.9471 |
| 0.0007 | 896.0 | 930944 | 0.5291 | 0.7995 | 0.8208 | 0.8100 | 0.9465 |
| 0.0007 | 897.0 | 931983 | 0.5391 | 0.8066 | 0.8151 | 0.8108 | 0.9465 |
| 0.0008 | 898.0 | 933022 | 0.5428 | 0.8003 | 0.8192 | 0.8096 | 0.9470 |
| 0.0008 | 899.0 | 934061 | 0.5463 | 0.7993 | 0.8212 | 0.8101 | 0.9460 |
| 0.0006 | 900.0 | 935100 | 0.5531 | 0.8021 | 0.8161 | 0.8091 | 0.9469 |
| 0.0006 | 901.0 | 936139 | 0.5665 | 0.8032 | 0.8116 | 0.8074 | 0.9460 |
| 0.0006 | 902.0 | 937178 | 0.5648 | 0.7992 | 0.8244 | 0.8116 | 0.9464 |
| 0.0006 | 903.0 | 938217 | 0.5461 | 0.8010 | 0.8185 | 0.8097 | 0.9462 |
| 0.0009 | 904.0 | 939256 | 0.5507 | 0.8011 | 0.8212 | 0.8110 | 0.9463 |
| 0.0007 | 905.0 | 940295 | 0.5507 | 0.7978 | 0.8187 | 0.8081 | 0.9455 |
| 0.0007 | 906.0 | 941334 | 0.5265 | 0.8027 | 0.8154 | 0.8090 | 0.9459 |
| 0.0007 | 907.0 | 942373 | 0.5403 | 0.7987 | 0.8230 | 0.8106 | 0.9463 |
| 0.0006 | 908.0 | 943412 | 0.5295 | 0.7980 | 0.8212 | 0.8094 | 0.9462 |
| 0.0006 | 909.0 | 944451 | 0.5407 | 0.7986 | 0.8203 | 0.8093 | 0.9464 |
| 0.001 | 910.0 | 945490 | 0.5218 | 0.8014 | 0.8199 | 0.8106 | 0.9470 |
| 0.0006 | 911.0 | 946529 | 0.5314 | 0.7997 | 0.8176 | 0.8085 | 0.9468 |
| 0.0008 | 912.0 | 947568 | 0.5052 | 0.7940 | 0.8224 | 0.8080 | 0.9464 |
| 0.0007 | 913.0 | 948607 | 0.5191 | 0.8031 | 0.8214 | 0.8121 | 0.9476 |
| 0.0007 | 914.0 | 949646 | 0.5203 | 0.8020 | 0.8248 | 0.8132 | 0.9477 |
| 0.0007 | 915.0 | 950685 | 0.5337 | 0.7969 | 0.8239 | 0.8102 | 0.9471 |
| 0.0007 | 916.0 | 951724 | 0.5192 | 0.7982 | 0.8194 | 0.8087 | 0.9469 |
| 0.0005 | 917.0 | 952763 | 0.5149 | 0.8025 | 0.8158 | 0.8091 | 0.9467 |
| 0.0005 | 918.0 | 953802 | 0.5294 | 0.7958 | 0.8248 | 0.8100 | 0.9473 |
| 0.0006 | 919.0 | 954841 | 0.5321 | 0.8008 | 0.8179 | 0.8093 | 0.9471 |
| 0.0006 | 920.0 | 955880 | 0.5382 | 0.8007 | 0.8190 | 0.8098 | 0.9464 |
| 0.0007 | 921.0 | 956919 | 0.5272 | 0.7991 | 0.8215 | 0.8102 | 0.9470 |
| 0.0004 | 922.0 | 957958 | 0.5427 | 0.8037 | 0.8160 | 0.8098 | 0.9463 |
| 0.0005 | 923.0 | 958997 | 0.5526 | 0.7973 | 0.8210 | 0.8090 | 0.9471 |
| 0.0005 | 924.0 | 960036 | 0.5417 | 0.8091 | 0.8142 | 0.8116 | 0.9469 |
| 0.0006 | 925.0 | 961075 | 0.5245 | 0.8036 | 0.8237 | 0.8135 | 0.9477 |
| 0.0005 | 926.0 | 962114 | 0.5453 | 0.8006 | 0.8269 | 0.8135 | 0.9477 |
| 0.0005 | 927.0 | 963153 | 0.5519 | 0.7970 | 0.8223 | 0.8094 | 0.9467 |
| 0.0008 | 928.0 | 964192 | 0.5326 | 0.8042 | 0.8226 | 0.8133 | 0.9467 |
| 0.0005 | 929.0 | 965231 | 0.5435 | 0.8004 | 0.8208 | 0.8105 | 0.9466 |
| 0.0004 | 930.0 | 966270 | 0.5637 | 0.7983 | 0.8223 | 0.8101 | 0.9468 |
| 0.0006 | 931.0 | 967309 | 0.5538 | 0.7988 | 0.8181 | 0.8084 | 0.9466 |
| 0.0005 | 932.0 | 968348 | 0.5329 | 0.8001 | 0.8197 | 0.8098 | 0.9465 |
| 0.0007 | 933.0 | 969387 | 0.5238 | 0.7967 | 0.8215 | 0.8089 | 0.9470 |
| 0.0003 | 934.0 | 970426 | 0.5384 | 0.8015 | 0.8203 | 0.8108 | 0.9472 |
| 0.0004 | 935.0 | 971465 | 0.5514 | 0.7968 | 0.8206 | 0.8086 | 0.9459 |
| 0.0004 | 936.0 | 972504 | 0.5423 | 0.8036 | 0.8253 | 0.8143 | 0.9472 |
| 0.0004 | 937.0 | 973543 | 0.5489 | 0.8016 | 0.8235 | 0.8124 | 0.9477 |
| 0.0003 | 938.0 | 974582 | 0.5641 | 0.7974 | 0.8221 | 0.8095 | 0.9462 |
| 0.0005 | 939.0 | 975621 | 0.5471 | 0.7982 | 0.8208 | 0.8094 | 0.9470 |
| 0.0004 | 940.0 | 976660 | 0.5547 | 0.7986 | 0.8190 | 0.8087 | 0.9465 |
| 0.0005 | 941.0 | 977699 | 0.5574 | 0.8030 | 0.8178 | 0.8103 | 0.9466 |
| 0.0006 | 942.0 | 978738 | 0.5537 | 0.8020 | 0.8194 | 0.8106 | 0.9469 |
| 0.0003 | 943.0 | 979777 | 0.5642 | 0.8029 | 0.8217 | 0.8122 | 0.9469 |
| 0.0004 | 944.0 | 980816 | 0.5575 | 0.7969 | 0.8237 | 0.8101 | 0.9468 |
| 0.0004 | 945.0 | 981855 | 0.5560 | 0.8017 | 0.8250 | 0.8132 | 0.9469 |
| 0.0007 | 946.0 | 982894 | 0.5472 | 0.8017 | 0.8278 | 0.8146 | 0.9472 |
| 0.0005 | 947.0 | 983933 | 0.5506 | 0.7993 | 0.8206 | 0.8098 | 0.9462 |
| 0.0004 | 948.0 | 984972 | 0.5486 | 0.7969 | 0.8275 | 0.8119 | 0.9458 |
| 0.0005 | 949.0 | 986011 | 0.5390 | 0.8019 | 0.8244 | 0.8130 | 0.9464 |
| 0.0004 | 950.0 | 987050 | 0.5459 | 0.8004 | 0.8260 | 0.8130 | 0.9472 |
| 0.0005 | 951.0 | 988089 | 0.5433 | 0.7983 | 0.8277 | 0.8127 | 0.9473 |
| 0.0004 | 952.0 | 989128 | 0.5575 | 0.7949 | 0.8241 | 0.8092 | 0.9460 |
| 0.0004 | 953.0 | 990167 | 0.5607 | 0.7953 | 0.8235 | 0.8092 | 0.9463 |
| 0.0003 | 954.0 | 991206 | 0.5659 | 0.8028 | 0.8228 | 0.8127 | 0.9471 |
| 0.0003 | 955.0 | 992245 | 0.5755 | 0.8015 | 0.8224 | 0.8118 | 0.9467 |
| 0.0004 | 956.0 | 993284 | 0.5596 | 0.8027 | 0.8226 | 0.8125 | 0.9469 |
| 0.0003 | 957.0 | 994323 | 0.5661 | 0.7988 | 0.8174 | 0.8080 | 0.9458 |
| 0.0003 | 958.0 | 995362 | 0.5691 | 0.7999 | 0.8242 | 0.8119 | 0.9469 |
| 0.0004 | 959.0 | 996401 | 0.5572 | 0.8028 | 0.8212 | 0.8119 | 0.9470 |
| 0.0002 | 960.0 | 997440 | 0.5651 | 0.8006 | 0.8284 | 0.8142 | 0.9471 |
| 0.0003 | 961.0 | 998479 | 0.5703 | 0.7989 | 0.8214 | 0.8100 | 0.9465 |
| 0.0002 | 962.0 | 999518 | 0.5838 | 0.8012 | 0.8223 | 0.8116 | 0.9464 |
| 0.0003 | 963.0 | 1000557 | 0.5831 | 0.8024 | 0.8206 | 0.8114 | 0.9468 |
| 0.0002 | 964.0 | 1001596 | 0.5915 | 0.8029 | 0.8199 | 0.8113 | 0.9469 |
| 0.0002 | 965.0 | 1002635 | 0.6012 | 0.8010 | 0.8212 | 0.8110 | 0.9467 |
| 0.0003 | 966.0 | 1003674 | 0.5766 | 0.8021 | 0.8239 | 0.8128 | 0.9471 |
| 0.0002 | 967.0 | 1004713 | 0.5802 | 0.8040 | 0.8208 | 0.8123 | 0.9470 |
| 0.0002 | 968.0 | 1005752 | 0.5982 | 0.8003 | 0.8226 | 0.8113 | 0.9468 |
| 0.0003 | 969.0 | 1006791 | 0.5910 | 0.8018 | 0.8230 | 0.8122 | 0.9472 |
| 0.0004 | 970.0 | 1007830 | 0.5951 | 0.7971 | 0.8251 | 0.8109 | 0.9471 |
| 0.0003 | 971.0 | 1008869 | 0.5956 | 0.8012 | 0.8232 | 0.8120 | 0.9464 |
| 0.0004 | 972.0 | 1009908 | 0.5847 | 0.7977 | 0.8242 | 0.8107 | 0.9468 |
| 0.0004 | 973.0 | 1010947 | 0.5894 | 0.8007 | 0.8203 | 0.8104 | 0.9468 |
| 0.0003 | 974.0 | 1011986 | 0.5888 | 0.8016 | 0.8214 | 0.8113 | 0.9469 |
| 0.0002 | 975.0 | 1013025 | 0.5888 | 0.7981 | 0.8181 | 0.8080 | 0.9465 |
| 0.0002 | 976.0 | 1014064 | 0.5864 | 0.7962 | 0.8237 | 0.8097 | 0.9462 |
| 0.0002 | 977.0 | 1015103 | 0.5911 | 0.7988 | 0.8217 | 0.8101 | 0.9465 |
| 0.0003 | 978.0 | 1016142 | 0.5955 | 0.7999 | 0.8176 | 0.8086 | 0.9462 |
| 0.0004 | 979.0 | 1017181 | 0.5881 | 0.8010 | 0.8176 | 0.8092 | 0.9462 |
| 0.0002 | 980.0 | 1018220 | 0.5868 | 0.7988 | 0.8206 | 0.8096 | 0.9467 |
| 0.0002 | 981.0 | 1019259 | 0.5834 | 0.8017 | 0.8206 | 0.8111 | 0.9470 |
| 0.0002 | 982.0 | 1020298 | 0.5843 | 0.7969 | 0.8237 | 0.8101 | 0.9469 |
| 0.0003 | 983.0 | 1021337 | 0.5830 | 0.7983 | 0.8246 | 0.8112 | 0.9470 |
| 0.0003 | 984.0 | 1022376 | 0.5822 | 0.7998 | 0.8230 | 0.8112 | 0.9468 |
| 0.0003 | 985.0 | 1023415 | 0.5826 | 0.8012 | 0.8221 | 0.8115 | 0.9468 |
| 0.0002 | 986.0 | 1024454 | 0.5883 | 0.8000 | 0.8217 | 0.8107 | 0.9467 |
| 0.0002 | 987.0 | 1025493 | 0.5917 | 0.7986 | 0.8242 | 0.8112 | 0.9465 |
| 0.0004 | 988.0 | 1026532 | 0.5864 | 0.8017 | 0.8214 | 0.8114 | 0.9466 |
| 0.0002 | 989.0 | 1027571 | 0.5880 | 0.7988 | 0.8235 | 0.8110 | 0.9466 |
| 0.0002 | 990.0 | 1028610 | 0.5873 | 0.7984 | 0.8230 | 0.8105 | 0.9471 |
| 0.0002 | 991.0 | 1029649 | 0.5871 | 0.8003 | 0.8235 | 0.8118 | 0.9469 |
| 0.0001 | 992.0 | 1030688 | 0.5874 | 0.7988 | 0.8224 | 0.8105 | 0.9466 |
| 0.0003 | 993.0 | 1031727 | 0.5878 | 0.8014 | 0.8217 | 0.8114 | 0.9468 |
| 0.0002 | 994.0 | 1032766 | 0.5920 | 0.7990 | 0.8217 | 0.8102 | 0.9464 |
| 0.0003 | 995.0 | 1033805 | 0.5899 | 0.8007 | 0.8206 | 0.8106 | 0.9466 |
| 0.0003 | 996.0 | 1034844 | 0.5906 | 0.8001 | 0.8205 | 0.8102 | 0.9467 |
| 0.0002 | 997.0 | 1035883 | 0.5907 | 0.8016 | 0.8214 | 0.8113 | 0.9467 |
| 0.0003 | 998.0 | 1036922 | 0.5888 | 0.8018 | 0.8210 | 0.8113 | 0.9468 |
| 0.0003 | 999.0 | 1037961 | 0.5893 | 0.8005 | 0.8206 | 0.8104 | 0.9467 |
| 0.0003 | 1000.0 | 1039000 | 0.5893 | 0.8007 | 0.8210 | 0.8107 | 0.9467 |
### Framework versions
- Transformers 4.29.2
- Pytorch 2.0.1+cu117
- Datasets 2.12.0
- Tokenizers 0.13.3
|
ssarae/dreambooth_jjangu3
|
ssarae
| 2023-10-06T03:06:45Z | 0 | 0 |
diffusers
|
[
"diffusers",
"tensorboard",
"stable-diffusion",
"stable-diffusion-diffusers",
"text-to-image",
"lora",
"base_model:runwayml/stable-diffusion-v1-5",
"base_model:adapter:runwayml/stable-diffusion-v1-5",
"license:creativeml-openrail-m",
"region:us"
] |
text-to-image
| 2023-09-26T01:19:47Z |
---
license: creativeml-openrail-m
base_model: runwayml/stable-diffusion-v1-5
instance_prompt: a photo of wkdrn jjangu
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
- diffusers
- lora
inference: true
---
# LoRA DreamBooth - ssarae/dreambooth_jjangu3
These are LoRA adaption weights for runwayml/stable-diffusion-v1-5. The weights were trained on a photo of wkdrn jjangu using [DreamBooth](https://dreambooth.github.io/). You can find some example images in the following.
LoRA for the text encoder was enabled: False.
|
GraydientPlatformAPI/realcreal
|
GraydientPlatformAPI
| 2023-10-06T03:01:00Z | 29 | 0 |
diffusers
|
[
"diffusers",
"text-to-image",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] |
text-to-image
| 2023-10-06T02:49:15Z |
---
library_name: diffusers
pipeline_tag: text-to-image
---
|
AustinMcMike/Stevie_v3
|
AustinMcMike
| 2023-10-06T02:55:51Z | 0 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-10-06T02:55:39Z |
---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- 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 0.5.0
|
Ehraim/results
|
Ehraim
| 2023-10-06T02:25:22Z | 0 | 0 | null |
[
"generated_from_trainer",
"base_model:tiiuae/falcon-7b-instruct",
"base_model:finetune:tiiuae/falcon-7b-instruct",
"license:apache-2.0",
"region:us"
] | null | 2023-09-29T20:57:31Z |
---
license: apache-2.0
base_model: tiiuae/falcon-7b-instruct
tags:
- generated_from_trainer
model-index:
- name: results
results: []
---
<!-- 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 [tiiuae/falcon-7b-instruct](https://huggingface.co/tiiuae/falcon-7b-instruct) on an unknown dataset.
It achieves the following results on the evaluation set:
- eval_loss: 1.1389
- eval_runtime: 3296.11
- eval_samples_per_second: 2.269
- eval_steps_per_second: 2.269
- epoch: 1.0
- step: 4362
## 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: 1
- 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: cosine
- lr_scheduler_warmup_ratio: 0.5
- num_epochs: 5
### Framework versions
- Transformers 4.34.0
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.14.0
|
ManiWavelabs/falcon-40b-trans
|
ManiWavelabs
| 2023-10-06T02:13:26Z | 0 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-10-06T00:28:03Z |
---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: True
- load_in_4bit: False
- 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: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
### Framework versions
- PEFT 0.6.0.dev0
|
garlandc/path_to_saved_model_rug
|
garlandc
| 2023-10-06T01:46:52Z | 31 | 0 |
diffusers
|
[
"diffusers",
"tensorboard",
"safetensors",
"stable-diffusion",
"stable-diffusion-diffusers",
"text-to-image",
"dreambooth",
"base_model:CompVis/stable-diffusion-v1-4",
"base_model:finetune:CompVis/stable-diffusion-v1-4",
"license:creativeml-openrail-m",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] |
text-to-image
| 2023-10-05T07:16:04Z |
---
license: creativeml-openrail-m
base_model: CompVis/stable-diffusion-v1-4
instance_prompt: a photo of sks rug
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
- diffusers
- dreambooth
inference: true
---
# DreamBooth - garlandc/path_to_saved_model_rug
This is a dreambooth model derived from CompVis/stable-diffusion-v1-4. The weights were trained on a photo of sks rug using [DreamBooth](https://dreambooth.github.io/).
You can find some example images in the following.
DreamBooth for the text encoder was enabled: False.
|
sonnguyen25/bloom-1b7_PROMPT_TUNING_CAUSAL_LM_2
|
sonnguyen25
| 2023-10-06T01:17:14Z | 1 | 0 |
peft
|
[
"peft",
"arxiv:1910.09700",
"base_model:bigscience/bloom-1b7",
"base_model:adapter:bigscience/bloom-1b7",
"region:us"
] | null | 2023-10-06T01:17:13Z |
---
library_name: peft
base_model: bigscience/bloom-1b7
---
# 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]
- **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 Data 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 Data 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]
## Training procedure
### Framework versions
- PEFT 0.6.0.dev0
|
grakshit/squad_a_r_train5_epoch_20
|
grakshit
| 2023-10-06T01:08:46Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"roberta",
"text-classification",
"generated_from_trainer",
"base_model:FacebookAI/roberta-base",
"base_model:finetune:FacebookAI/roberta-base",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
text-classification
| 2023-10-05T18:54:58Z |
---
license: mit
base_model: roberta-base
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: squad_a_r_train5_epoch_20
results: []
---
<!-- 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. -->
# squad_a_r_train5_epoch_20
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: 0.0008
- Accuracy: 0.9994
## 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: 20
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 0.6038 | 1.0 | 1188 | 0.5918 | 0.6938 |
| 0.5954 | 2.0 | 2376 | 0.6000 | 0.6957 |
| 0.5713 | 3.0 | 3564 | 0.5453 | 0.7342 |
| 0.5346 | 4.0 | 4752 | 0.4763 | 0.7841 |
| 0.4804 | 5.0 | 5940 | 0.3846 | 0.8298 |
| 0.4057 | 6.0 | 7128 | 0.2845 | 0.8835 |
| 0.3303 | 7.0 | 8316 | 0.1906 | 0.9259 |
| 0.2739 | 8.0 | 9504 | 0.1240 | 0.9564 |
| 0.2121 | 9.0 | 10692 | 0.0994 | 0.9714 |
| 0.1821 | 10.0 | 11880 | 0.0958 | 0.9756 |
| 0.1639 | 11.0 | 13068 | 0.0596 | 0.9853 |
| 0.1361 | 12.0 | 14256 | 0.0407 | 0.9895 |
| 0.1116 | 13.0 | 15444 | 0.0318 | 0.9924 |
| 0.0872 | 14.0 | 16632 | 0.0216 | 0.9954 |
| 0.072 | 15.0 | 17820 | 0.0125 | 0.9975 |
| 0.0648 | 16.0 | 19008 | 0.0167 | 0.9960 |
| 0.0537 | 17.0 | 20196 | 0.0126 | 0.9979 |
| 0.0404 | 18.0 | 21384 | 0.0045 | 0.9992 |
| 0.027 | 19.0 | 22572 | 0.0033 | 0.9987 |
| 0.0274 | 20.0 | 23760 | 0.0008 | 0.9994 |
### Framework versions
- Transformers 4.34.0
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.14.0
|
Guapeton/distilbert-base-uncased-finetuned-squad
|
Guapeton
| 2023-10-06T01:03:41Z | 116 | 0 |
transformers
|
[
"transformers",
"pytorch",
"distilbert",
"question-answering",
"generated_from_trainer",
"dataset:squad",
"base_model:distilbert/distilbert-base-uncased",
"base_model:finetune:distilbert/distilbert-base-uncased",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
question-answering
| 2023-10-03T13:31:29Z |
---
license: apache-2.0
base_model: distilbert-base-uncased
tags:
- generated_from_trainer
datasets:
- squad
model-index:
- name: distilbert-base-uncased-finetuned-squad
results: []
---
<!-- 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. -->
# distilbert-base-uncased-finetuned-squad
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the squad 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: 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
### Framework versions
- Transformers 4.34.0
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.14.0
|
alikanakar/xlsr-syntesized-turkish-4-hour-llr
|
alikanakar
| 2023-10-06T00:57:07Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"wav2vec2",
"automatic-speech-recognition",
"generated_from_trainer",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
automatic-speech-recognition
| 2023-10-05T22:02:43Z |
---
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: xlsr-syntesized-turkish-4-hour-llr
results: []
---
<!-- 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. -->
# xlsr-syntesized-turkish-4-hour-llr
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 3.7455
- Wer: 1.0
## 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: 1e-05
- train_batch_size: 2
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 200
- num_epochs: 20
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:-----:|:----:|:---------------:|:---:|
| 9.2067 | 0.52 | 100 | 7.7166 | 1.0 |
| 4.7666 | 1.04 | 200 | 4.5059 | 1.0 |
| 3.6173 | 1.56 | 300 | 3.6445 | 1.0 |
| 3.3747 | 2.08 | 400 | 3.4531 | 1.0 |
| 3.2351 | 2.6 | 500 | 3.3077 | 1.0 |
| 3.16 | 3.12 | 600 | 3.1965 | 1.0 |
| 3.092 | 3.65 | 700 | 3.1473 | 1.0 |
| 3.0051 | 4.17 | 800 | 3.1395 | 1.0 |
| 2.785 | 4.69 | 900 | 3.2007 | 1.0 |
| 2.3783 | 5.21 | 1000 | 3.1456 | 1.0 |
| 2.0074 | 5.73 | 1100 | 3.0829 | 1.0 |
| 1.767 | 6.25 | 1200 | 3.0563 | 1.0 |
| 1.4704 | 6.77 | 1300 | 3.0522 | 1.0 |
| 1.2749 | 7.29 | 1400 | 3.1219 | 1.0 |
| 1.12 | 7.81 | 1500 | 3.1256 | 1.0 |
| 1.0151 | 8.33 | 1600 | 3.1973 | 1.0 |
| 0.9098 | 8.85 | 1700 | 3.2901 | 1.0 |
| 0.8451 | 9.38 | 1800 | 3.2970 | 1.0 |
| 0.793 | 9.9 | 1900 | 3.3592 | 1.0 |
| 0.7518 | 10.42 | 2000 | 3.4237 | 1.0 |
| 0.6994 | 10.94 | 2100 | 3.4695 | 1.0 |
| 0.6686 | 11.46 | 2200 | 3.5052 | 1.0 |
| 0.6558 | 11.98 | 2300 | 3.5431 | 1.0 |
| 0.6247 | 12.5 | 2400 | 3.6077 | 1.0 |
| 0.615 | 13.02 | 2500 | 3.6611 | 1.0 |
| 0.5889 | 13.54 | 2600 | 3.5684 | 1.0 |
| 0.5702 | 14.06 | 2700 | 3.5891 | 1.0 |
| 0.5552 | 14.58 | 2800 | 3.6027 | 1.0 |
| 0.5367 | 15.1 | 2900 | 3.6398 | 1.0 |
| 0.5322 | 15.62 | 3000 | 3.6480 | 1.0 |
| 0.5051 | 16.15 | 3100 | 3.6087 | 1.0 |
| 0.5063 | 16.67 | 3200 | 3.6777 | 1.0 |
| 0.4869 | 17.19 | 3300 | 3.6848 | 1.0 |
| 0.4945 | 17.71 | 3400 | 3.6886 | 1.0 |
| 0.4848 | 18.23 | 3500 | 3.7217 | 1.0 |
| 0.489 | 18.75 | 3600 | 3.7360 | 1.0 |
| 0.4811 | 19.27 | 3700 | 3.7442 | 1.0 |
| 0.4665 | 19.79 | 3800 | 3.7455 | 1.0 |
### Framework versions
- Transformers 4.26.0
- Pytorch 2.0.1+cu118
- Datasets 2.9.0
- Tokenizers 0.13.3
|
soBeauty/20231005-2-bert-base-multilingual-cased-new
|
soBeauty
| 2023-10-06T00:34:06Z | 195 | 0 |
transformers
|
[
"transformers",
"pytorch",
"bert",
"fill-mask",
"generated_from_trainer",
"base_model:google-bert/bert-base-multilingual-cased",
"base_model:finetune:google-bert/bert-base-multilingual-cased",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
fill-mask
| 2023-10-05T19:20:03Z |
---
license: apache-2.0
base_model: bert-base-multilingual-cased
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: 20231005-2-bert-base-multilingual-cased-new
results: []
---
<!-- 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. -->
# 20231005-2-bert-base-multilingual-cased-new
This model is a fine-tuned version of [bert-base-multilingual-cased](https://huggingface.co/bert-base-multilingual-cased) on an unknown dataset.
It achieves the following results on the evaluation set:
- Accuracy: 0.6479
- Loss: 1.6281
## 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: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 20
### Training results
| Training Loss | Epoch | Step | Accuracy | Validation Loss |
|:-------------:|:-----:|:----:|:--------:|:---------------:|
| 2.8918 | 1.82 | 200 | 0.4038 | 2.4903 |
| 2.3882 | 3.64 | 400 | 0.4978 | 2.2096 |
| 2.1785 | 5.45 | 600 | 0.4934 | 1.9995 |
| 2.0686 | 7.27 | 800 | 0.5386 | 2.0103 |
| 1.9575 | 9.09 | 1000 | 0.5243 | 2.1245 |
| 1.8828 | 10.91 | 1200 | 0.5932 | 1.7872 |
| 1.8112 | 12.73 | 1400 | 0.6204 | 1.4894 |
| 1.7449 | 14.55 | 1600 | 0.5978 | 1.7968 |
| 1.6916 | 16.36 | 1800 | 0.6253 | 1.6254 |
| 1.6763 | 18.18 | 2000 | 0.6377 | 1.4638 |
| 1.6459 | 20.0 | 2200 | 0.6479 | 1.6281 |
### Framework versions
- Transformers 4.34.0
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.14.0
|
TheBloke/llama-2-7B-Arguments-GGUF
|
TheBloke
| 2023-10-06T00:17:49Z | 125 | 5 |
transformers
|
[
"transformers",
"gguf",
"llama",
"base_model:cris177/llama-2-7b-Arguments",
"base_model:quantized:cris177/llama-2-7b-Arguments",
"license:llama2",
"region:us"
] | null | 2023-10-05T22:39:35Z |
---
base_model: cris177/llama-2-7b-Arguments
inference: false
license: llama2
model_creator: Cristian Desivo
model_name: Llama 2 7B Arguments
model_type: llama
prompt_template: '<s>[INST] {prompt}
[/INST]
'
quantized_by: TheBloke
---
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Llama 2 7B Arguments - GGUF
- Model creator: [Cristian Desivo](https://huggingface.co/cris177)
- Original model: [Llama 2 7B Arguments](https://huggingface.co/cris177/llama-2-7b-Arguments)
<!-- description start -->
## Description
This repo contains GGUF format model files for [Cristian Desivo's Llama 2 7B Arguments](https://huggingface.co/cris177/llama-2-7b-Arguments).
<!-- description end -->
<!-- README_GGUF.md-about-gguf start -->
### About GGUF
GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.
Here is an incomplate list of clients and libraries that are known to support GGUF:
* [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option.
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.
* [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.
* [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration.
* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection.
* [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.
* [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server.
* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.
* [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use.
<!-- README_GGUF.md-about-gguf end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/llama-2-7B-Arguments-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/llama-2-7B-Arguments-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/llama-2-7B-Arguments-GGUF)
* [Cristian Desivo's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/cris177/llama-2-7b-Arguments)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: Llama-2-Prompt
```
<s>[INST] {prompt}
[/INST]
```
<!-- prompt-template end -->
<!-- compatibility_gguf start -->
## Compatibility
These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221)
They are also compatible with many third party UIs and libraries - please see the list at the top of this README.
## Explanation of quantisation methods
<details>
<summary>Click to see details</summary>
The new methods available are:
* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
Refer to the Provided Files table below to see what files use which methods, and how.
</details>
<!-- compatibility_gguf end -->
<!-- README_GGUF.md-provided-files start -->
## Provided files
| Name | Quant method | Bits | Size | Max RAM required | Use case |
| ---- | ---- | ---- | ---- | ---- | ----- |
| [llama-2-7b-arguments.Q2_K.gguf](https://huggingface.co/TheBloke/llama-2-7B-Arguments-GGUF/blob/main/llama-2-7b-arguments.Q2_K.gguf) | Q2_K | 2 | 2.83 GB| 5.33 GB | smallest, significant quality loss - not recommended for most purposes |
| [llama-2-7b-arguments.Q3_K_S.gguf](https://huggingface.co/TheBloke/llama-2-7B-Arguments-GGUF/blob/main/llama-2-7b-arguments.Q3_K_S.gguf) | Q3_K_S | 3 | 2.95 GB| 5.45 GB | very small, high quality loss |
| [llama-2-7b-arguments.Q3_K_M.gguf](https://huggingface.co/TheBloke/llama-2-7B-Arguments-GGUF/blob/main/llama-2-7b-arguments.Q3_K_M.gguf) | Q3_K_M | 3 | 3.30 GB| 5.80 GB | very small, high quality loss |
| [llama-2-7b-arguments.Q3_K_L.gguf](https://huggingface.co/TheBloke/llama-2-7B-Arguments-GGUF/blob/main/llama-2-7b-arguments.Q3_K_L.gguf) | Q3_K_L | 3 | 3.60 GB| 6.10 GB | small, substantial quality loss |
| [llama-2-7b-arguments.Q4_0.gguf](https://huggingface.co/TheBloke/llama-2-7B-Arguments-GGUF/blob/main/llama-2-7b-arguments.Q4_0.gguf) | Q4_0 | 4 | 3.83 GB| 6.33 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
| [llama-2-7b-arguments.Q4_K_S.gguf](https://huggingface.co/TheBloke/llama-2-7B-Arguments-GGUF/blob/main/llama-2-7b-arguments.Q4_K_S.gguf) | Q4_K_S | 4 | 3.86 GB| 6.36 GB | small, greater quality loss |
| [llama-2-7b-arguments.Q4_K_M.gguf](https://huggingface.co/TheBloke/llama-2-7B-Arguments-GGUF/blob/main/llama-2-7b-arguments.Q4_K_M.gguf) | Q4_K_M | 4 | 4.08 GB| 6.58 GB | medium, balanced quality - recommended |
| [llama-2-7b-arguments.Q5_0.gguf](https://huggingface.co/TheBloke/llama-2-7B-Arguments-GGUF/blob/main/llama-2-7b-arguments.Q5_0.gguf) | Q5_0 | 5 | 4.65 GB| 7.15 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
| [llama-2-7b-arguments.Q5_K_S.gguf](https://huggingface.co/TheBloke/llama-2-7B-Arguments-GGUF/blob/main/llama-2-7b-arguments.Q5_K_S.gguf) | Q5_K_S | 5 | 4.65 GB| 7.15 GB | large, low quality loss - recommended |
| [llama-2-7b-arguments.Q5_K_M.gguf](https://huggingface.co/TheBloke/llama-2-7B-Arguments-GGUF/blob/main/llama-2-7b-arguments.Q5_K_M.gguf) | Q5_K_M | 5 | 4.78 GB| 7.28 GB | large, very low quality loss - recommended |
| [llama-2-7b-arguments.Q6_K.gguf](https://huggingface.co/TheBloke/llama-2-7B-Arguments-GGUF/blob/main/llama-2-7b-arguments.Q6_K.gguf) | Q6_K | 6 | 5.53 GB| 8.03 GB | very large, extremely low quality loss |
| [llama-2-7b-arguments.Q8_0.gguf](https://huggingface.co/TheBloke/llama-2-7B-Arguments-GGUF/blob/main/llama-2-7b-arguments.Q8_0.gguf) | Q8_0 | 8 | 7.16 GB| 9.66 GB | very large, extremely low quality loss - not recommended |
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
<!-- README_GGUF.md-provided-files end -->
<!-- README_GGUF.md-how-to-download start -->
## How to download GGUF files
**Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.
The following clients/libraries will automatically download models for you, providing a list of available models to choose from:
- LM Studio
- LoLLMS Web UI
- Faraday.dev
### In `text-generation-webui`
Under Download Model, you can enter the model repo: TheBloke/llama-2-7B-Arguments-GGUF and below it, a specific filename to download, such as: llama-2-7b-arguments.Q4_K_M.gguf.
Then click Download.
### On the command line, including multiple files at once
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
Then you can download any individual model file to the current directory, at high speed, with a command like this:
```shell
huggingface-cli download TheBloke/llama-2-7B-Arguments-GGUF llama-2-7b-arguments.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
<details>
<summary>More advanced huggingface-cli download usage</summary>
You can also download multiple files at once with a pattern:
```shell
huggingface-cli download TheBloke/llama-2-7B-Arguments-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'
```
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/llama-2-7B-Arguments-GGUF llama-2-7b-arguments.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
</details>
<!-- README_GGUF.md-how-to-download end -->
<!-- README_GGUF.md-how-to-run start -->
## Example `llama.cpp` command
Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
```shell
./main -ngl 32 -m llama-2-7b-arguments.Q4_K_M.gguf --color -c 4096 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "<s>[INST] {prompt}\n[/INST]"
```
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
Change `-c 4096` to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically.
If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`
For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md)
## How to run in `text-generation-webui`
Further instructions here: [text-generation-webui/docs/llama.cpp.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/llama.cpp.md).
## How to run from Python code
You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries.
### How to load this model in Python code, using ctransformers
#### First install the package
Run one of the following commands, according to your system:
```shell
# Base ctransformers with no GPU acceleration
pip install ctransformers
# Or with CUDA GPU acceleration
pip install ctransformers[cuda]
# Or with AMD ROCm GPU acceleration (Linux only)
CT_HIPBLAS=1 pip install ctransformers --no-binary ctransformers
# Or with Metal GPU acceleration for macOS systems only
CT_METAL=1 pip install ctransformers --no-binary ctransformers
```
#### Simple ctransformers example code
```python
from ctransformers import AutoModelForCausalLM
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
llm = AutoModelForCausalLM.from_pretrained("TheBloke/llama-2-7B-Arguments-GGUF", model_file="llama-2-7b-arguments.Q4_K_M.gguf", model_type="llama", gpu_layers=50)
print(llm("AI is going to"))
```
## How to use with LangChain
Here are guides on using llama-cpp-python and ctransformers with LangChain:
* [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)
* [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)
<!-- README_GGUF.md-how-to-run end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Pierre Kircher, Stanislav Ovsiannikov, Michael Levine, Eugene Pentland, Andrey, 준교 김, Randy H, Fred von Graf, Artur Olbinski, Caitlyn Gatomon, terasurfer, Jeff Scroggin, James Bentley, Vadim, Gabriel Puliatti, Harry Royden McLaughlin, Sean Connelly, Dan Guido, Edmond Seymore, Alicia Loh, subjectnull, AzureBlack, Manuel Alberto Morcote, Thomas Belote, Lone Striker, Chris Smitley, Vitor Caleffi, Johann-Peter Hartmann, Clay Pascal, biorpg, Brandon Frisco, sidney chen, transmissions 11, Pedro Madruga, jinyuan sun, Ajan Kanaga, Emad Mostaque, Trenton Dambrowitz, Jonathan Leane, Iucharbius, usrbinkat, vamX, George Stoitzev, Luke Pendergrass, theTransient, Olakabola, Swaroop Kallakuri, Cap'n Zoog, Brandon Phillips, Michael Dempsey, Nikolai Manek, danny, Matthew Berman, Gabriel Tamborski, alfie_i, Raymond Fosdick, Tom X Nguyen, Raven Klaugh, LangChain4j, Magnesian, Illia Dulskyi, David Ziegler, Mano Prime, Luis Javier Navarrete Lozano, Erik Bjäreholt, 阿明, Nathan Dryer, Alex, Rainer Wilmers, zynix, TL, Joseph William Delisle, John Villwock, Nathan LeClaire, Willem Michiel, Joguhyik, GodLy, OG, Alps Aficionado, Jeffrey Morgan, ReadyPlayerEmma, Tiffany J. Kim, Sebastain Graf, Spencer Kim, Michael Davis, webtim, Talal Aujan, knownsqashed, John Detwiler, Imad Khwaja, Deo Leter, Jerry Meng, Elijah Stavena, Rooh Singh, Pieter, SuperWojo, Alexandros Triantafyllidis, Stephen Murray, Ai Maven, ya boyyy, Enrico Ros, Ken Nordquist, Deep Realms, Nicholas, Spiking Neurons AB, Elle, Will Dee, Jack West, RoA, Luke @flexchar, Viktor Bowallius, Derek Yates, Subspace Studios, jjj, Toran Billups, Asp the Wyvern, Fen Risland, Ilya, NimbleBox.ai, Chadd, Nitin Borwankar, Emre, Mandus, Leonard Tan, Kalila, K, Trailburnt, S_X, Cory Kujawski
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
<!-- original-model-card start -->
# Original model card: Cristian Desivo's Llama 2 7B Arguments
This model aims to detect and analyze casual arguments.
Model template:
```
<s>[INST] {prompt}
[/INST]
```
Example:
```
`<`s`>`[INST] Analize the following argument, identifying premises, conclusion, type of argument, and argument validity:
If officer smith found a broken window at the crime scene then the arson occurred on elm street, and officer smith found a broken window at the crime scene, hence the arson occurred on elm street.
[/INST] Premise 1: If officer smith found a broken window at the crime scene then the arson occurred on elm street Premise 2: Officer smith found a broken window at the crime scene Conclusion: The arson occurred on Elm Street Type of argument: modus ponen Validity: True `<`/s`>`
```
It was trained on my dataset cris177/Arguments (https://huggingface.co/datasets/cris177/Arguments)
<!-- original-model-card end -->
|
mostafaashahin/voiced
|
mostafaashahin
| 2023-10-06T00:17:08Z | 5 | 0 |
generic
|
[
"generic",
"pytorch",
"wav2vec2",
"automatic-speech-recognition",
"region:us"
] |
automatic-speech-recognition
| 2023-10-05T11:02:48Z |
---
tags:
- automatic-speech-recognition
library_name: generic
---
# Automatic Speech Recognition repository template
This is a template repository for Automatic Speech Recognition to support generic inference with Hugging Face Hub generic Inference API. There are two required steps:
1. Specify the requirements by defining a `requirements.txt` file.
2. Implement the `pipeline.py` `__init__` and `__call__` methods. These methods are called by the Inference API. The `__init__` method should load the model and preload all the elements needed for inference (model, processors, tokenizers, etc.). This is only called once. The `__call__` method performs the actual inference. Make sure to follow the same input/output specifications defined in the template for the pipeline to work.
Example repos
* https://huggingface.co/osanseviero/pyctcdecode_asr
## How to start
First create a repo in https://hf.co/new.
Then clone this template and push it to your repo.
```
git clone https://huggingface.co/templates/automatic-speech-recognition
cd automatic-speech-recognition
git remote set-url origin https://huggingface.co/$YOUR_USER/$YOUR_REPO_NAME
git push --force
```
|
ManiWavelabs/falconwave40b
|
ManiWavelabs
| 2023-10-06T00:15:25Z | 2 | 0 |
peft
|
[
"peft",
"pytorch",
"falcon",
"custom_code",
"8-bit",
"bitsandbytes",
"region:us"
] | null | 2023-10-05T22:53:57Z |
---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: True
- load_in_4bit: False
- 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: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: True
- load_in_4bit: False
- 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: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
### Framework versions
- PEFT 0.6.0.dev0
- PEFT 0.6.0.dev0
|
daekeun-ml/ko-trocr-base-nsmc-news-chatbot
|
daekeun-ml
| 2023-10-05T23:59:58Z | 349 | 10 |
transformers
|
[
"transformers",
"pytorch",
"safetensors",
"vision-encoder-decoder",
"image-text-to-text",
"trocr",
"image-to-text",
"ko",
"license:mit",
"endpoints_compatible",
"region:us"
] |
image-to-text
| 2022-11-22T00:05:16Z |
---
language:
- ko
tags:
- trocr
- image-to-text
license: mit
metrics:
- wer
- cer
widget:
- src: https://raw.githubusercontent.com/aws-samples/sm-kornlp/main/trocr/sample_imgs/random_2.jpg
example_title: 랜덤 문장 1
- src: https://raw.githubusercontent.com/aws-samples/sm-kornlp/main/trocr/sample_imgs/random_6.jpg
example_title: 랜덤 문장 2
- src: https://raw.githubusercontent.com/aws-samples/sm-kornlp/main/trocr/sample_imgs/chatbot_3.jpg
example_title: 챗봇 1
- src: https://raw.githubusercontent.com/aws-samples/sm-kornlp/main/trocr/sample_imgs/chatbot_5.jpg
example_title: 챗봇 2
- src: https://raw.githubusercontent.com/aws-samples/sm-kornlp/main/trocr/sample_imgs/news_1.jpg
example_title: 뉴스 1
- src: https://raw.githubusercontent.com/aws-samples/sm-kornlp/main/trocr/sample_imgs/news_3.jpg
example_title: 뉴스 2
- src: https://raw.githubusercontent.com/aws-samples/sm-kornlp/main/trocr/sample_imgs/nsmc_1.jpg
example_title: 영화 리뷰 1
- src: https://raw.githubusercontent.com/aws-samples/sm-kornlp/main/trocr/sample_imgs/nsmc_2.jpg
example_title: 영화 리뷰 2
---
# TrOCR for Korean Language (PoC)
## Overview
TrOCR has not yet released a multilingual model including Korean, so we trained a Korean model for PoC purpose. Based on this model, it is recommended to collect more data to additionally train the 1st stage or perform fine-tuning as the 2nd stage.
## Collecting data
### Text data
We created training data by processing three types of datasets.
- News summarization dataset: https://huggingface.co/datasets/daekeun-ml/naver-news-summarization-ko
- Naver Movie Sentiment Classification: https://github.com/e9t/nsmc
- Chatbot dataset: https://github.com/songys/Chatbot_data
For efficient data collection, each sentence was separated by a sentence separator library (Kiwi Python wrapper; https://github.com/bab2min/kiwipiepy), and as a result, 637,401 samples were collected.
### Image Data
Image data was generated with TextRecognitionDataGenerator (https://github.com/Belval/TextRecognitionDataGenerator) introduced in the TrOCR paper.
Below is a code snippet for generating images.
```shell
python3 ./trdg/run.py -i ocr_dataset_poc.txt -w 5 -t {num_cores} -f 64 -l ko -c {num_samples} -na 2 --output_dir {dataset_dir}
```
## Training
### Base model
The encoder model used `facebook/deit-base-distilled-patch16-384` and the decoder model used `klue/roberta-base`. It is easier than training by starting weights from `microsoft/trocr-base-stage1`.
### Parameters
We used heuristic parameters without separate hyperparameter tuning.
- learning_rate = 4e-5
- epochs = 25
- fp16 = True
- max_length = 64
## Usage
### inference.py
```python
from transformers import TrOCRProcessor, VisionEncoderDecoderModel, AutoTokenizer
import requests
from io import BytesIO
from PIL import Image
processor = TrOCRProcessor.from_pretrained("microsoft/trocr-base-handwritten")
model = VisionEncoderDecoderModel.from_pretrained("daekeun-ml/ko-trocr-base-nsmc-news-chatbot")
tokenizer = AutoTokenizer.from_pretrained("daekeun-ml/ko-trocr-base-nsmc-news-chatbot")
url = "https://raw.githubusercontent.com/aws-samples/sm-kornlp/main/trocr/sample_imgs/news_1.jpg"
response = requests.get(url)
img = Image.open(BytesIO(response.content))
pixel_values = processor(img, return_tensors="pt").pixel_values
generated_ids = model.generate(pixel_values, max_length=64)
generated_text = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
print(generated_text)
```
All the code required for data collection and model training has been published on the author's Github.
- https://github.com/daekeun-ml/sm-kornlp-usecases/tree/main/trocr
|
yzzhong/ppo-SnowballTarget
|
yzzhong
| 2023-10-05T23:46:36Z | 5 | 0 |
ml-agents
|
[
"ml-agents",
"tensorboard",
"onnx",
"SnowballTarget",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-SnowballTarget",
"region:us"
] |
reinforcement-learning
| 2023-10-05T23:46:33Z |
---
library_name: ml-agents
tags:
- SnowballTarget
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-SnowballTarget
---
# **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: yzzhong/ppo-SnowballTarget
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
scionk/xlm-roberta-base-finetuned-panx-en
|
scionk
| 2023-10-05T23:44:06Z | 103 | 0 |
transformers
|
[
"transformers",
"pytorch",
"xlm-roberta",
"token-classification",
"generated_from_trainer",
"dataset:xtreme",
"base_model:FacebookAI/xlm-roberta-base",
"base_model:finetune:FacebookAI/xlm-roberta-base",
"license:mit",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] |
token-classification
| 2023-09-25T06:12:08Z |
---
license: mit
base_model: xlm-roberta-base
tags:
- generated_from_trainer
datasets:
- xtreme
metrics:
- f1
model-index:
- name: xlm-roberta-base-finetuned-panx-en
results:
- task:
name: Token Classification
type: token-classification
dataset:
name: xtreme
type: xtreme
config: PAN-X.en
split: validation
args: PAN-X.en
metrics:
- name: F1
type: f1
value: 0.6893854748603352
---
<!-- 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. -->
# xlm-roberta-base-finetuned-panx-en
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the xtreme dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4013
- F1: 0.6894
## 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: 24
- eval_batch_size: 24
- 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 | F1 |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 1.1107 | 1.0 | 50 | 0.5691 | 0.5586 |
| 0.5078 | 2.0 | 100 | 0.4297 | 0.6579 |
| 0.3574 | 3.0 | 150 | 0.4013 | 0.6894 |
### Framework versions
- Transformers 4.33.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.5
- Tokenizers 0.13.3
|
Numlock/svn-opt-2
|
Numlock
| 2023-10-05T23:42:55Z | 1 | 0 |
peft
|
[
"peft",
"arxiv:1910.09700",
"base_model:facebook/opt-6.7b",
"base_model:adapter:facebook/opt-6.7b",
"region:us"
] | null | 2023-10-05T23:42:05Z |
---
library_name: peft
base_model: facebook/opt-6.7b
---
# 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]
- **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 Data 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 Data 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]
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- load_in_8bit: True
- load_in_4bit: False
- 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: fp4
- bnb_4bit_use_double_quant: False
- bnb_4bit_compute_dtype: float32
### Framework versions
- PEFT 0.6.0.dev0
|
Yntec/Classic
|
Yntec
| 2023-10-05T23:40:47Z | 520 | 3 |
diffusers
|
[
"diffusers",
"safetensors",
"Anime",
"Cartoon",
"Zovya",
"DucHaiten",
"stable-diffusion",
"stable-diffusion-diffusers",
"text-to-image",
"license:creativeml-openrail-m",
"autotrain_compatible",
"endpoints_compatible",
"diffusers:StableDiffusionPipeline",
"region:us"
] |
text-to-image
| 2023-10-05T22:42:35Z |
---
license: creativeml-openrail-m
library_name: diffusers
pipeline_tag: text-to-image
tags:
- Anime
- Cartoon
- Zovya
- DucHaiten
- stable-diffusion
- stable-diffusion-diffusers
- diffusers
- text-to-image
---
# Classic
A mix of DucHaitenClassicAnime and CartoonStyleClassic by Zovya to make a model of the best from western and japanese animation! It uses the MoistMixV2 VAE.
(click for larger)
Sample and prompt:
videogames, robert jordan pepperoni pizza, josephine wall winner, hidari, roll20 illumination, sitting Pretty CUTE girl, radiant light, gorgeous hair, DETAILED EYES, Magazine ad, iconic, 1943, Cartoon, sharp focus, 4k, towel. comic art on canvas by kyoani and watched and ROSSDRAWS. elementary
Original pages:
https://civitai.com/models/33030/cartoon-style-classic
https://civitai.com/models/8542?modelVersionId=16168 (DucHaitenClassicAnime)
# Recipe
- SuperMerger Weight sum MBW 0,0,0,0,0,0,0,0,0,0,0,0,1,1,1,1,1,1,1,1,1,1,1,1,1,1
Model A: DucHaitenClassicAnime
Model B: CartoonStyleClassic
Output: Classic
- Bake MoistMixV2 VAE:
Output: ClassicVAE
|
snorkelai/med-flan-t5-large
|
snorkelai
| 2023-10-05T23:29:14Z | 8 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"t5",
"text2text-generation",
"simplification",
"generated_from_trainer",
"es",
"dataset:lcampillos/CLARA-MeD",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] |
text2text-generation
| 2023-10-05T23:28:29Z |
---
license: apache-2.0
tags:
- simplification
- generated_from_trainer
metrics:
- rouge
- sari
model-index:
- name: flan-t5-large-clara-med
results: []
datasets:
- lcampillos/CLARA-MeD
language:
- es
---
<!-- 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. -->
# flan-t5-large-clara-med
This model is a fine-tuned version of [google/flan-t5-large](https://huggingface.co/google/flan-t5-large) on the [CLARA-MeD](https://huggingface.co/lcampillos/CLARA-MeD) dataset.
It achieves the following results on the evaluation set:
- Loss: 1.0898
- Rouge1: 28.9888
- Rouge2: 16.3801
- Rougel: 27.4186
- Rougelsum: 27.4043
- sari: 39.1731
## 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: 5.6e-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: 5
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|
| No log | 1.0 | 380 | 1.1948 | 28.2861 | 15.6461 | 26.7126 | 26.7389 |
| No log | 2.0 | 760 | 1.1361 | 28.3528 | 15.8519 | 26.8151 | 26.8069 |
| 1.3561 | 3.0 | 1140 | 1.1051 | 29.6216 | 16.8227 | 28.0662 | 28.0613 |
| 1.3561 | 4.0 | 1520 | 1.0915 | 29.3603 | 16.5008 | 27.7915 | 27.7761 |
| 1.0939 | 5.0 | 1900 | 1.0898 | 28.9888 | 16.3801 | 27.4186 | 27.4043 |
### Framework versions
- Transformers 4.25.1
- Pytorch 1.13.0
- Datasets 2.8.0
- Tokenizers 0.12.1
|
TheBloke/airoboros-m-7B-3.0-GGUF
|
TheBloke
| 2023-10-05T23:27:14Z | 168 | 4 |
transformers
|
[
"transformers",
"gguf",
"mistral",
"dataset:jondurbin/airoboros-3.0",
"base_model:jondurbin/airoboros-m-7b-3.0",
"base_model:quantized:jondurbin/airoboros-m-7b-3.0",
"license:apache-2.0",
"region:us"
] | null | 2023-10-05T22:47:54Z |
---
base_model: jondurbin/airoboros-m-7b-3.0
datasets:
- jondurbin/airoboros-3.0
inference: false
license: apache-2.0
model_creator: Jon Durbin
model_name: Airoboros M 7B 3.0
model_type: mistral
prompt_template: '[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
'
quantized_by: TheBloke
---
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Airoboros M 7B 3.0 - GGUF
- Model creator: [Jon Durbin](https://huggingface.co/jondurbin)
- Original model: [Airoboros M 7B 3.0](https://huggingface.co/jondurbin/airoboros-m-7b-3.0)
<!-- description start -->
## Description
This repo contains GGUF format model files for [Jon Durbin's Airoboros M 7B 3.0](https://huggingface.co/jondurbin/airoboros-m-7b-3.0).
<!-- description end -->
<!-- README_GGUF.md-about-gguf start -->
### About GGUF
GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.
Here is an incomplate list of clients and libraries that are known to support GGUF:
* [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option.
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.
* [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.
* [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration.
* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection.
* [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.
* [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server.
* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.
* [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use.
<!-- README_GGUF.md-about-gguf end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/airoboros-m-7B-3.0-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/airoboros-m-7B-3.0-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/airoboros-m-7B-3.0-GGUF)
* [Jon Durbin's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/jondurbin/airoboros-m-7b-3.0)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: Llama-2-Uncensored
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
```
<!-- prompt-template end -->
<!-- compatibility_gguf start -->
## Compatibility
These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221)
They are also compatible with many third party UIs and libraries - please see the list at the top of this README.
## Explanation of quantisation methods
<details>
<summary>Click to see details</summary>
The new methods available are:
* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
Refer to the Provided Files table below to see what files use which methods, and how.
</details>
<!-- compatibility_gguf end -->
<!-- README_GGUF.md-provided-files start -->
## Provided files
| Name | Quant method | Bits | Size | Max RAM required | Use case |
| ---- | ---- | ---- | ---- | ---- | ----- |
| [airoboros-m-7b-3.0.Q2_K.gguf](https://huggingface.co/TheBloke/airoboros-m-7B-3.0-GGUF/blob/main/airoboros-m-7b-3.0.Q2_K.gguf) | Q2_K | 2 | 3.08 GB| 5.58 GB | smallest, significant quality loss - not recommended for most purposes |
| [airoboros-m-7b-3.0.Q3_K_S.gguf](https://huggingface.co/TheBloke/airoboros-m-7B-3.0-GGUF/blob/main/airoboros-m-7b-3.0.Q3_K_S.gguf) | Q3_K_S | 3 | 3.16 GB| 5.66 GB | very small, high quality loss |
| [airoboros-m-7b-3.0.Q3_K_M.gguf](https://huggingface.co/TheBloke/airoboros-m-7B-3.0-GGUF/blob/main/airoboros-m-7b-3.0.Q3_K_M.gguf) | Q3_K_M | 3 | 3.52 GB| 6.02 GB | very small, high quality loss |
| [airoboros-m-7b-3.0.Q3_K_L.gguf](https://huggingface.co/TheBloke/airoboros-m-7B-3.0-GGUF/blob/main/airoboros-m-7b-3.0.Q3_K_L.gguf) | Q3_K_L | 3 | 3.82 GB| 6.32 GB | small, substantial quality loss |
| [airoboros-m-7b-3.0.Q4_0.gguf](https://huggingface.co/TheBloke/airoboros-m-7B-3.0-GGUF/blob/main/airoboros-m-7b-3.0.Q4_0.gguf) | Q4_0 | 4 | 4.11 GB| 6.61 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
| [airoboros-m-7b-3.0.Q4_K_S.gguf](https://huggingface.co/TheBloke/airoboros-m-7B-3.0-GGUF/blob/main/airoboros-m-7b-3.0.Q4_K_S.gguf) | Q4_K_S | 4 | 4.14 GB| 6.64 GB | small, greater quality loss |
| [airoboros-m-7b-3.0.Q4_K_M.gguf](https://huggingface.co/TheBloke/airoboros-m-7B-3.0-GGUF/blob/main/airoboros-m-7b-3.0.Q4_K_M.gguf) | Q4_K_M | 4 | 4.37 GB| 6.87 GB | medium, balanced quality - recommended |
| [airoboros-m-7b-3.0.Q5_0.gguf](https://huggingface.co/TheBloke/airoboros-m-7B-3.0-GGUF/blob/main/airoboros-m-7b-3.0.Q5_0.gguf) | Q5_0 | 5 | 5.00 GB| 7.50 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
| [airoboros-m-7b-3.0.Q5_K_S.gguf](https://huggingface.co/TheBloke/airoboros-m-7B-3.0-GGUF/blob/main/airoboros-m-7b-3.0.Q5_K_S.gguf) | Q5_K_S | 5 | 5.00 GB| 7.50 GB | large, low quality loss - recommended |
| [airoboros-m-7b-3.0.Q5_K_M.gguf](https://huggingface.co/TheBloke/airoboros-m-7B-3.0-GGUF/blob/main/airoboros-m-7b-3.0.Q5_K_M.gguf) | Q5_K_M | 5 | 5.13 GB| 7.63 GB | large, very low quality loss - recommended |
| [airoboros-m-7b-3.0.Q6_K.gguf](https://huggingface.co/TheBloke/airoboros-m-7B-3.0-GGUF/blob/main/airoboros-m-7b-3.0.Q6_K.gguf) | Q6_K | 6 | 5.94 GB| 8.44 GB | very large, extremely low quality loss |
| [airoboros-m-7b-3.0.Q8_0.gguf](https://huggingface.co/TheBloke/airoboros-m-7B-3.0-GGUF/blob/main/airoboros-m-7b-3.0.Q8_0.gguf) | Q8_0 | 8 | 7.70 GB| 10.20 GB | very large, extremely low quality loss - not recommended |
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
<!-- README_GGUF.md-provided-files end -->
<!-- README_GGUF.md-how-to-download start -->
## How to download GGUF files
**Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.
The following clients/libraries will automatically download models for you, providing a list of available models to choose from:
- LM Studio
- LoLLMS Web UI
- Faraday.dev
### In `text-generation-webui`
Under Download Model, you can enter the model repo: TheBloke/airoboros-m-7B-3.0-GGUF and below it, a specific filename to download, such as: airoboros-m-7b-3.0.Q4_K_M.gguf.
Then click Download.
### On the command line, including multiple files at once
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
Then you can download any individual model file to the current directory, at high speed, with a command like this:
```shell
huggingface-cli download TheBloke/airoboros-m-7B-3.0-GGUF airoboros-m-7b-3.0.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
<details>
<summary>More advanced huggingface-cli download usage</summary>
You can also download multiple files at once with a pattern:
```shell
huggingface-cli download TheBloke/airoboros-m-7B-3.0-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'
```
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/airoboros-m-7B-3.0-GGUF airoboros-m-7b-3.0.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
</details>
<!-- README_GGUF.md-how-to-download end -->
<!-- README_GGUF.md-how-to-run start -->
## Example `llama.cpp` command
Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
```shell
./main -ngl 32 -m airoboros-m-7b-3.0.Q4_K_M.gguf --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "[INST] <<SYS>>\nYou are a help, unbiased, uncensored assistant.\n<</SYS>\n\n{prompt} [/INST]"
```
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
Change `-c 2048` to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically.
If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`
For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md)
## How to run in `text-generation-webui`
Further instructions here: [text-generation-webui/docs/llama.cpp.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/llama.cpp.md).
## How to run from Python code
You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries.
### How to load this model in Python code, using ctransformers
#### First install the package
Run one of the following commands, according to your system:
```shell
# Base ctransformers with no GPU acceleration
pip install ctransformers
# Or with CUDA GPU acceleration
pip install ctransformers[cuda]
# Or with AMD ROCm GPU acceleration (Linux only)
CT_HIPBLAS=1 pip install ctransformers --no-binary ctransformers
# Or with Metal GPU acceleration for macOS systems only
CT_METAL=1 pip install ctransformers --no-binary ctransformers
```
#### Simple ctransformers example code
```python
from ctransformers import AutoModelForCausalLM
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
llm = AutoModelForCausalLM.from_pretrained("TheBloke/airoboros-m-7B-3.0-GGUF", model_file="airoboros-m-7b-3.0.Q4_K_M.gguf", model_type="mistral", gpu_layers=50)
print(llm("AI is going to"))
```
## How to use with LangChain
Here are guides on using llama-cpp-python and ctransformers with LangChain:
* [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)
* [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)
<!-- README_GGUF.md-how-to-run end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Pierre Kircher, Stanislav Ovsiannikov, Michael Levine, Eugene Pentland, Andrey, 준교 김, Randy H, Fred von Graf, Artur Olbinski, Caitlyn Gatomon, terasurfer, Jeff Scroggin, James Bentley, Vadim, Gabriel Puliatti, Harry Royden McLaughlin, Sean Connelly, Dan Guido, Edmond Seymore, Alicia Loh, subjectnull, AzureBlack, Manuel Alberto Morcote, Thomas Belote, Lone Striker, Chris Smitley, Vitor Caleffi, Johann-Peter Hartmann, Clay Pascal, biorpg, Brandon Frisco, sidney chen, transmissions 11, Pedro Madruga, jinyuan sun, Ajan Kanaga, Emad Mostaque, Trenton Dambrowitz, Jonathan Leane, Iucharbius, usrbinkat, vamX, George Stoitzev, Luke Pendergrass, theTransient, Olakabola, Swaroop Kallakuri, Cap'n Zoog, Brandon Phillips, Michael Dempsey, Nikolai Manek, danny, Matthew Berman, Gabriel Tamborski, alfie_i, Raymond Fosdick, Tom X Nguyen, Raven Klaugh, LangChain4j, Magnesian, Illia Dulskyi, David Ziegler, Mano Prime, Luis Javier Navarrete Lozano, Erik Bjäreholt, 阿明, Nathan Dryer, Alex, Rainer Wilmers, zynix, TL, Joseph William Delisle, John Villwock, Nathan LeClaire, Willem Michiel, Joguhyik, GodLy, OG, Alps Aficionado, Jeffrey Morgan, ReadyPlayerEmma, Tiffany J. Kim, Sebastain Graf, Spencer Kim, Michael Davis, webtim, Talal Aujan, knownsqashed, John Detwiler, Imad Khwaja, Deo Leter, Jerry Meng, Elijah Stavena, Rooh Singh, Pieter, SuperWojo, Alexandros Triantafyllidis, Stephen Murray, Ai Maven, ya boyyy, Enrico Ros, Ken Nordquist, Deep Realms, Nicholas, Spiking Neurons AB, Elle, Will Dee, Jack West, RoA, Luke @flexchar, Viktor Bowallius, Derek Yates, Subspace Studios, jjj, Toran Billups, Asp the Wyvern, Fen Risland, Ilya, NimbleBox.ai, Chadd, Nitin Borwankar, Emre, Mandus, Leonard Tan, Kalila, K, Trailburnt, S_X, Cory Kujawski
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
<!-- original-model-card start -->
# Original model card: Jon Durbin's Airoboros M 7B 3.0
### Overview
Another experimental model, using mostly sythetic data generated by [airoboros](https://github.com/jondurbin/airoboros)
#### IMPORTANT NOTE - llama-2 chat format!!!
The airobors 3.0 models have been updated to use llama-2 chat format, rather than vicuna style user/assistant!
This is a breaking change, although most inference systems support llama-2 chat templating.
See prompt format below for all of the details.
#### Highlights
A model built on [airoboros-3.0 dataset](https://hf.co/datasets/jondurbin/airoboros-3.0), which is a continuation of [airoboros-2.2.1 dataset](https://hf.co/datasets/jondurbin/airoboros-2.2.1) with the following extras:
* MathJSON - this is not yet part of the code directly, but can be easily parsed and executed, as a way to perform deterministic calculations. LLMs suck at math, so this does problem formulation and let's the work be done by traditional operations.
* Human-generated RP multi-turn conversation - provided by an anon, this is a human created RP dataset, meant to assist with multi-turn coherency and chat experience.
This is a fairly general purpose model, but focuses heavily on instruction following, rather than casual chat/roleplay.
Huge thank you to the folks over at [a16z](https://a16z.com/) for sponsoring the costs associated with building models and associated tools!
### Prompt format
The prompt format is llama-2 chat.
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
```
For multi-turn, the prompt format is as follows:
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt 0} [/INST] {response 0} </s><s>[INST] {prompt 1} [/INST] {response 1} </s><s>...[INST] {prompt N} [/INST]
```
Here is a simple python script that can help generate the prompt:
```python
def get_chat_prompt(message, chat_history, system_prompt):
texts = [f'[INST] <<SYS>>\n{system_prompt}\n<</SYS>>\n\n']
do_strip = False
for user_input, response in chat_history:
user_input = user_input.strip() if do_strip else user_input
do_strip = True
texts.append(f'{user_input} [/INST] {response.strip()} </s><s>[INST] ')
message = message.strip() if do_strip else message
texts.append(f'{message} [/INST] ')
return ''.join(texts)
def get_prompt(item):
"""
`item` is expected to be in chatml format, e.g.
{
"id": 1,
"conversations": [
{
"role": "system",
"content": "You are a helpful, unbiased, uncensored assistant."
},
{
"role": "user",
"content": "Hi!"
},
{
"role": "assistant",
"content": "Hello, how are you today?"
},
{
"role": "user",
"content": "Just peachy, and you?"
}
]
}
"""
system_prompt = share_gpt_item['conversations'][0]['value']
offset = 1
if share_gpt_item['conversations'][0]['role'] != 'system':
system_prompt = 'You are an unbiased, uncensored, helpful assistant.'
offset = 0
chat_history = [
(share_gpt_item['conversations'][idx]['content'], share_gpt_item['conversations'][idx + 1]['content'])
for idx in range(offset, len(item['conversations']) - 1, 2)
]
message = item['conversations'][-1]['content']
return get_chat_prompt(message, chat_history, system_prompt)
```
### Helpful usage tips
#### MathJSON
Ask mathematical questions, prefixed with "Create a MathJSON solution to the following:", and you should get output with a formula that can be executed by https://cortexjs.io/compute-engine/
I also made a (really quite hacky, incomplete, semi-functional) python version that works for most expressions, see [mathjson.py](https://github.com/jondurbin/airoboros/blob/mathjson/airoboros/mathjson.py)
__Use a very low temperature!__
Example:
```
[INST] <<SYS>>
You are an assistant with vast knowledge in all things math.
<</SYS>
Create a MathJSON solution to the following: Calculate the area of a circle with radius of 17.2456cm. Include your reasoning. [/INST]
```
Output:
```
The area (A) of a circle is given by the formula A = πr², where r is the radius of the circle. Since the radius is given in centimeters, we need to convert it to meters first. We know that 1 meter equals 100 centimeters. Therefore, we can represent the radius as: r = 17.2456 cm * 100/1.
Once we have the radius in meters, we can calculate the area of the circle using the formula A = πr².
Solution as MathJSON:
<mathjson>
[
"Multiply",
"Pi",
[
"Power",
17.2456,
2
]
]
</mathjson>
```
You can then validate the JSON between `<mathjson>` and `</mathjson>`, then pass the parsed JSON to compute-engine JS or the `evaluate` function in mathjson.py to calculate the response.
#### Context obedient question answering
By obedient, I mean the model was trained to ignore what it thinks it knows, and uses the context to answer the question. The model was also tuned to limit the values to the provided context as much as possible to reduce hallucinations.
The format for a closed-context prompt is as follows:
```
BEGININPUT
BEGINCONTEXT
[key0: value0]
[key1: value1]
... other metdata ...
ENDCONTEXT
[insert your text blocks here]
ENDINPUT
[add as many other blocks, in the exact same format]
BEGININSTRUCTION
[insert your instruction(s). The model was tuned with single questions, paragraph format, lists, etc.]
ENDINSTRUCTION
```
It's also helpful to add "Don't make up answers if you don't know." to your instruction block to make sure if the context is completely unrelated it doesn't make something up.
*The __only__ prompts that need this closed context formating are closed-context instructions. Normal questions/instructions do not!*
I know it's a bit verbose and annoying, but after much trial and error, using these explicit delimiters helps the model understand where to find the responses and how to associate specific sources with it.
- `BEGININPUT` - denotes a new input block
- `BEGINCONTEXT` - denotes the block of context (metadata key/value pairs) to associate with the current input block
- `ENDCONTEXT` - denotes the end of the metadata block for the current input
- [text] - Insert whatever text you want for the input block, as many paragraphs as can fit in the context.
- `ENDINPUT` - denotes the end of the current input block
- [repeat as many input blocks in this format as you want]
- `BEGININSTRUCTION` - denotes the start of the list (or one) instruction(s) to respond to for all of the input blocks above.
- [instruction(s)]
- `ENDINSTRUCTION` - denotes the end of instruction set
It sometimes works without `ENDINSTRUCTION`, but by explicitly including that in the prompt, the model better understands that all of the instructions in the block should be responded to.
__Use a very low temperature!__
Here's a trivial, but important example to prove the point:
```
BEGININPUT
BEGINCONTEXT
date: 2021-01-01
url: https://web.site/123
ENDCONTEXT
In a shocking turn of events, blueberries are now green, but will be sticking with the same name.
ENDINPUT
BEGININSTRUCTION
What color are bluberries? Source?
ENDINSTRUCTION
```
And the response:
```
Blueberries are now green.
Source:
date: 2021-01-01
url: https://web.site/123
```
#### Summarization
500 samples have been included from [this dataset](https://huggingface.co/datasets/mattpscott/airoboros-summarization), using the same format as contextual question answering, for example:
```
BEGININPUT
{text to summarize}
ENDINPUT
BEGININSTRUCTION
Summarize the input in around 130 words.
ENDINSTRUCTION
```
#### Getting longer responses
You can use a few techniques to get longer responses.
Detailed prompts, with explicit instruction for word count:
```
Please compose a narrative set in the heart of an ancient library, steeped in the scent of old parchment and ink. The protagonist should be a young scholar who is dedicated to studying the art of storytelling and its evolution throughout history. In her pursuit of knowledge, she stumbles upon a forgotten tome that seems to possess an unusual aura. This book has the ability to bring stories to life, literally manifesting characters and scenarios from within its pages into reality.
The main character must navigate through various epochs of storytelling - from oral traditions of tribal societies, through medieval minstrels' tales, to modern-day digital narratives - as they come alive around her. Each era presents its unique challenges and lessons about the power and impact of stories on human civilization.
One such character could be a sentient quill pen, who was once used by renowned authors of yesteryears and now holds their wisdom and experiences. It becomes her mentor, guiding her through this journey with witty remarks and insightful commentary.
Ensure that your tale encapsulates the thrill of adventure, the beauty of learning, and the profound connection between humans and their stories. All characters involved should be non-human entities. Feel free to explore creative liberties but maintain the mentioned elements.
Your response should be approximately 2300 words.
```
Or, a simpler example:
```
Please create a long, detailed story about a dragon in an old growth forest who, for some reason, begins speaking the words of the source code of linux.
```
There are a few examples of next chapter completion as well, e.g.:
```
Write the next chapter of a historical fiction novel set in Paris during the 20th century.
Here's a summary of the previous chapter:
In the vibrant city of Paris, amid the tumultuous changes of the 20th century, our protagonist Margot, an aspiring fashion designer, has just secured an apprenticeship at a prestigious couture house. She meets Lucien, a charming journalist who covers the fashion industry. Together they navigate the ever-changing world of fashion and society, uncovering secrets that reveal the intricate links between style, politics, and culture. As the chapter concludes, they decide to delve deeper into the hidden corners of the fashion world to unravel its mysteries.
Requirements for the next chapter:
1. Character Development of Margot and Lucien:
- Margot's Evolution: Unfold more about Margot's past, her dreams of revolutionizing fashion, and her struggle to establish herself in a male-dominated industry. Illustrate her growing expertise, innovative ideas, and increasing dependence on Lucien.
- Lucien's Complexity: Introduce uncertainties surrounding Lucien's background and real motives. Increase suspense by suggesting undisclosed information he possesses, while also highlighting his wit and perceptiveness.
2. Exploration of Paris and the Couture House:
- Paris: Elaborate their journey through the bustling streets of Paris, including encounters with iconic figures, social unrest, and relics from different eras of French history.
- The Couture House: Expand on the grandeur of the couture house they work in, filled with artistic masterpieces, intense competition, and cryptic notes hinting at a scandalous past.
3. Emergence of the Subplot: The Lost Collection:
- Discovery: Have Margot and Lucien stumble upon a secret vault containing a lost collection designed before World War II, raising new questions about the previous owner and the influence of war on fashion.
- Revelation: Capture their shock as they realize the designs were plagiarized, the potential repercussions, and the opportunities it presents for Margot's career.
- Twist: End with a twist that suggests there are other stolen collections across Paris, setting up their new mission.
Your response should be approximately 650 words.
```
#### Coding
You can ask for fairly complex coding instructions with multiple criteria, e.g.:
```
Create a python application with the following requirements:
- Asyncio FastAPI webserver
- ping endpoint that returns the current date in JSON format
- file upload endpoint, which calculates the file's sha256 checksum, and checks postgres to deduplicate
```
Or inline criteria:
```
Write a multi-threaded TCP server in C that accepts a "GET [key]" input and "SET [key] [value]" input, and uses a binary tree to get and store the input values.
```
You can also optionally add a single space and "PLAINFORMAT" at the end of your prompt to avoid backticks, explanations, etc. and just print the code, e.g.:
```
Write a websocket application in node.js. PLAINFORMAT
```
#### Agent/function calling
The dataset includes many examples of function/args generation based on input criteria. This is somewhat similar to the OpenAI function calling, but the output is either JSON or YAML.
Example prompt:
```
As an AI assistant, please select the most suitable function and parameters from the list of available functions below, based on the user's input. Provide your response in JSON format.
Input: I want to know how many times 'Python' is mentioned in my text file.
Available functions:
file_analytics:
description: This tool performs various operations on a text file.
params:
action: The operation we want to perform on the data, such as "count_occurrences", "find_line", etc.
filters:
keyword: The word or phrase we want to search for.
```
Response:
```json
{
"function": "file_analytics",
"params": {
"action": "count_occurrences",
"filters": {
"keyword": "Python"
}
}
}
```
#### Chain-of-thought
You can ask for several possible responses to a given problem, with a ranking and final answer selection.
Example prompt:
```
A group of people decide to play a round-robin tournament where every player plays every other player exactly once. If a total of 45 games were played, how many players participated in the tournament? Offer several potential answers, rank them according to mathematical logic, and pick the most feasible one.
```
Example response:
```
Solution 1:
In a round-robin tournament, each player plays every other player exactly once. This means that the total number of games played is equal to the combination of all players taken two at a time (since a game involves two players).
The formula for combinations is nC2 = n(n-1)/2, where n is the total number of players. We know that nC2 equals 45 (the total number of games), so we can set up the equation as follows:
n(n-1)/2 = 45
Solving this equation gives us n=10.
Final answer: There were 10 players in the tournament.
Solution 2:
Another way to approach this problem is by using the concept of triangular numbers. Triangular numbers are a sequence of numbers where each number can be represented as a triangle with dots. The nth triangular number is given by the sum of the first n natural numbers.
If we consider each game as a dot and each player as a side of the triangle, then the total number of games (45) would correspond to the 9th triangular number because 1+2+3+4+5+6+7+8+9=45. However, since each side of the triangle represents a player, and there's one more player than sides in our model (because the last player has no one left to play against), we need to add one more to get the total number of players.
Final answer: There were 10 players in the tournament.
Solution 3:
We could also use trial and error to solve this problem. Starting from a small number of players, we calculate how many games would be played in a round-robin tournament until we reach 45 games. For example, with 2 players, only 1 game is played; with 3 players, 3 games are played; with 4 players, 6 games are played, and so on. Continuing this process, we find that with 10 players, 45 games are played.
Final answer: There were 10 players in the tournament.
Ranking of solutions from best to worst:
1. Solution 1: This solution uses direct mathematical reasoning based on the properties of a round-robin tournament and solves an equation to find the answer.
2. Solution 2: This solution uses the concept of triangular numbers which requires additional steps and knowledge beyond the immediate problem context, but still arrives at the correct answer.
3. Solution 3: This solution relies on trial and error, which is less efficient and reliable than the other methods, but it eventually gets to the right answer.
Best and final answer: There were 10 players in the tournament.
```
#### reWOO style execution planning
The model now supports execution planning for complex instructions that would require making use of several tools. The output is just the plan, you must implement a mechanism to parse the output and actually call the functions!
Example prompt:
```
Please construct a systematic plan to generate an optimal response to the user instruction, utilizing a set of provided tools. Each plan will correspond to an evidence value, which will be the output of one of the available functions given an input string
that could be the user's question, one or more prior evidence values, or a combination of both.
Here are the tools available to be called:
DuckDuckGo[input]: Tool that enables the user to search for information using the DuckDuckGo search engine. This tool is helpful in locating a suitable list of sites that may contain the answer to the user's query. It does not directly answer the question but finds a list of sites that might have the answer.
TextScraper[input]: Load one or more websites from the input string containing newline delimited links, where input is one or more links, and produces plain text output containing the content of the links.
HyperlinkExtractor[input]: Extract links from plain text and produces a plain text, newline delimited response of links.
KnowledgeModel[input]: Question answering language model, particularly useful in answering questions based on an input passage of text. The input must be a text question that references an :evidence[n]: variable, e.g. What color is the cat, given :evidence1:?
The output should be in format:
Plan: [first action to take based on input question]
:evidence0: = [function to call with input parameter]
Plan: [next action to take, based on result of :evidence0:]
:evidence1: = [next function to call and input parameter, which may include reference to previous evidence, e.g. "Given context :evidence0"]
...
Answer: [:evidence[n]: containing the final answer.]
Question: Who won the Nobel Prize in Literature in 2020 and what is their most famous work?
```
Response:
```
Plan: Begin by conducting a web search to find out who won the Nobel Prize in Literature in 2020.
:evidence0: = DuckDuckGo[Nobel Prize in Literature 2020 winner]
Plan: Utilize the KnowledgeModel to interpret the search results and find the name of the laureate.
:evidence1: = KnowledgeModel[Who won the Nobel Prize in Literature in 2020, given :evidence0:?]
Plan: Conduct another web search to find the most famous work of the identified laureate.
:evidence2: = DuckDuckGo[Most famous work of :evidence1:]
Plan: Extract the relevant links from the DuckDuckGo search results for a more focused search.
:evidence3: = HyperlinkExtractor[:evidence2:]
Plan: Use the TextScraper tool to extract information from the relevant links.
:evidence4: = TextScraper[:evidence3:]
Plan: Finally, utilize the KnowledgeModel to identify and summarize the most famous work of the laureate from the extracted information.
:evidence5: = KnowledgeModel[What is the most famous work of :evidence1:, given :evidence4:?]
Answer: :evidence5:
```
For this to be useful, you'd have to parse the output plan text, and implement/call each of the functions. This is just pseudo-code, completely untested off the top of my head, and obviously would requiring full implementation + hardening:
```python
import re
import requests
def inject_context(input_text, **context):
for ref in set(re.findall(r"(:evidence[0-9]+:)", input_text, re.I)):
input_text = input_text.replace(ref, context.get(ref, ""))
return input_text
def duckduckgo(input_text, **context):
search_string = inject_context(input_text, **context)
... search via duck duck go using search_string
... return text content
def link_extractor(input_text, **context):
input_text = inject_context(input_text, **context)
return "\n".join(list(set(re.findall(r"(https?://[^\s]+?\.?)", input_text, re.I))))
def scrape(input_text, **context):
input_text = inject_context(input_text, **context)
text = []
for link in input_text.splitlines():
text.append(requests.get(link).text)
return "\n".join(text)
def infer(input_text, **context)
prompt = inject_context(input_text, **context)
... call model with prompt, return output
def parse_plan(plan):
method_map = {
"DuckDuckGo": duckduckgo,
"HyperlinkExtractor": link_extractor,
"KnowledgeModel": infer,
"TextScraper": scrape,
}
context = {}
for line in plan.strip().splitlines():
if line.startswith("Plan:"):
print(line)
continue
parts = re.match("^(:evidence[0-9]+:)\s*=\s*([^\[]+])(\[.*\])\s$", line, re.I)
if not parts:
if line.startswith("Answer: "):
return context.get(line.split(" ")[-1].strip(), "Answer couldn't be generated...")
raise RuntimeError("bad format: " + line)
context[parts.group(1)] = method_map[parts.group(2)](parts.group(3), **context)
```
### Contribute
If you're interested in new functionality, particularly a new "instructor" type to generate a specific type of training data,
take a look at the dataset generation tool repo: https://github.com/jondurbin/airoboros and either make a PR or open an issue with details.
To help me with the OpenAI/compute costs:
- https://bmc.link/jondurbin
- ETH 0xce914eAFC2fe52FdceE59565Dd92c06f776fcb11
- BTC bc1qdwuth4vlg8x37ggntlxu5cjfwgmdy5zaa7pswf
### Licence and usage restrictions
The airoboros 3.0 models are built on top of multiple base models, each with their own license/restrictions.
The models with `-l2` in the name have a custom Meta license:
- See the [meta-license/LICENSE.txt](meta-license/LICENSE.txt) file attached for the original license provided by Meta.
- See also [meta-license/USE_POLICY.md](meta-license/USE_POLICY.md) and [meta-license/Responsible-Use-Guide.pdf](meta-license/Responsible-Use-Guide.pdf), also provided by Meta.
The models with `-m-` are mistral-7b (apache 2.0)
The model with `-3b` uses Stability AI, which as a `cc-by-sa-4.0` license.
The fine-tuning data was mostly generated by OpenAI API calls to gpt-4, via [airoboros](https://github.com/jondurbin/airoboros)
The ToS for OpenAI API usage has a clause preventing the output from being used to train a model that __competes__ with OpenAI
- what does *compete* actually mean here?
- these small open source models will not produce output anywhere near the quality of gpt-4, or even gpt-3.5, so I can't imagine this could credibly be considered competing in the first place
- if someone else uses the dataset to do the same, they wouldn't necessarily be violating the ToS because they didn't call the API, so I don't know how that works
- the training data used in essentially all large language models includes a significant amount of copyrighted or otherwise non-permissive licensing in the first place
- other work using the self-instruct method, e.g. the original here: https://github.com/yizhongw/self-instruct released the data and model as apache-2
I am purposingly leaving this license ambiguous (other than the fact you must comply with the Meta original license for llama-2) because I am not a lawyer and refuse to attempt to interpret all of the terms accordingly.
Your best bet is probably to avoid using this commercially due to the OpenAI API usage.
Either way, by using this model, you agree to completely indemnify me.
<!-- original-model-card end -->
|
TheBloke/airoboros-l2-7B-3.0-GPTQ
|
TheBloke
| 2023-10-05T23:26:36Z | 18 | 0 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"dataset:jondurbin/airoboros-3.0",
"base_model:jondurbin/airoboros-l2-7b-3.0",
"base_model:quantized:jondurbin/airoboros-l2-7b-3.0",
"license:llama2",
"autotrain_compatible",
"text-generation-inference",
"4-bit",
"gptq",
"region:us"
] |
text-generation
| 2023-10-05T23:09:32Z |
---
base_model: jondurbin/airoboros-l2-7b-3.0
datasets:
- jondurbin/airoboros-3.0
inference: false
license: llama2
model_creator: Jon Durbin
model_name: Airoboros L2 7B 3.0
model_type: llama
prompt_template: '[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
'
quantized_by: TheBloke
---
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Airoboros L2 7B 3.0 - GPTQ
- Model creator: [Jon Durbin](https://huggingface.co/jondurbin)
- Original model: [Airoboros L2 7B 3.0](https://huggingface.co/jondurbin/airoboros-l2-7b-3.0)
<!-- description start -->
## Description
This repo contains GPTQ model files for [Jon Durbin's Airoboros L2 7B 3.0](https://huggingface.co/jondurbin/airoboros-l2-7b-3.0).
Multiple GPTQ parameter permutations are provided; see Provided Files below for details of the options provided, their parameters, and the software used to create them.
<!-- description end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GGUF)
* [Jon Durbin's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/jondurbin/airoboros-l2-7b-3.0)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: Llama-2-Uncensored
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
```
<!-- prompt-template end -->
<!-- README_GPTQ.md-provided-files start -->
## Provided files, and GPTQ parameters
Multiple quantisation parameters are provided, to allow you to choose the best one for your hardware and requirements.
Each separate quant is in a different branch. See below for instructions on fetching from different branches.
Most GPTQ files are made with AutoGPTQ. Mistral models are currently made with Transformers.
<details>
<summary>Explanation of GPTQ parameters</summary>
- Bits: The bit size of the quantised model.
- GS: GPTQ group size. Higher numbers use less VRAM, but have lower quantisation accuracy. "None" is the lowest possible value.
- Act Order: True or False. Also known as `desc_act`. True results in better quantisation accuracy. Some GPTQ clients have had issues with models that use Act Order plus Group Size, but this is generally resolved now.
- Damp %: A GPTQ parameter that affects how samples are processed for quantisation. 0.01 is default, but 0.1 results in slightly better accuracy.
- GPTQ dataset: The calibration dataset used during quantisation. Using a dataset more appropriate to the model's training can improve quantisation accuracy. Note that the GPTQ calibration dataset is not the same as the dataset used to train the model - please refer to the original model repo for details of the training dataset(s).
- Sequence Length: The length of the dataset sequences used for quantisation. Ideally this is the same as the model sequence length. For some very long sequence models (16+K), a lower sequence length may have to be used. Note that a lower sequence length does not limit the sequence length of the quantised model. It only impacts the quantisation accuracy on longer inference sequences.
- ExLlama Compatibility: Whether this file can be loaded with ExLlama, which currently only supports Llama models in 4-bit.
</details>
| Branch | Bits | GS | Act Order | Damp % | GPTQ Dataset | Seq Len | Size | ExLlama | Desc |
| ------ | ---- | -- | --------- | ------ | ------------ | ------- | ---- | ------- | ---- |
| [main](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GPTQ/tree/main) | 4 | 128 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 3.90 GB | Yes | 4-bit, with Act Order and group size 128g. Uses even less VRAM than 64g, but with slightly lower accuracy. |
| [gptq-4bit-32g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GPTQ/tree/gptq-4bit-32g-actorder_True) | 4 | 32 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 4.28 GB | Yes | 4-bit, with Act Order and group size 32g. Gives highest possible inference quality, with maximum VRAM usage. |
| [gptq-8bit--1g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GPTQ/tree/gptq-8bit--1g-actorder_True) | 8 | None | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 7.01 GB | No | 8-bit, with Act Order. No group size, to lower VRAM requirements. |
| [gptq-8bit-128g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GPTQ/tree/gptq-8bit-128g-actorder_True) | 8 | 128 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 7.16 GB | No | 8-bit, with group size 128g for higher inference quality and with Act Order for even higher accuracy. |
| [gptq-8bit-32g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GPTQ/tree/gptq-8bit-32g-actorder_True) | 8 | 32 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 7.62 GB | No | 8-bit, with group size 32g and Act Order for maximum inference quality. |
| [gptq-4bit-64g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GPTQ/tree/gptq-4bit-64g-actorder_True) | 4 | 64 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 4.02 GB | Yes | 4-bit, with Act Order and group size 64g. Uses less VRAM than 32g, but with slightly lower accuracy. |
<!-- README_GPTQ.md-provided-files end -->
<!-- README_GPTQ.md-download-from-branches start -->
## How to download, including from branches
### In text-generation-webui
To download from the `main` branch, enter `TheBloke/airoboros-l2-7B-3.0-GPTQ` in the "Download model" box.
To download from another branch, add `:branchname` to the end of the download name, eg `TheBloke/airoboros-l2-7B-3.0-GPTQ:gptq-4bit-32g-actorder_True`
### From the command line
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
To download the `main` branch to a folder called `airoboros-l2-7B-3.0-GPTQ`:
```shell
mkdir airoboros-l2-7B-3.0-GPTQ
huggingface-cli download TheBloke/airoboros-l2-7B-3.0-GPTQ --local-dir airoboros-l2-7B-3.0-GPTQ --local-dir-use-symlinks False
```
To download from a different branch, add the `--revision` parameter:
```shell
mkdir airoboros-l2-7B-3.0-GPTQ
huggingface-cli download TheBloke/airoboros-l2-7B-3.0-GPTQ --revision gptq-4bit-32g-actorder_True --local-dir airoboros-l2-7B-3.0-GPTQ --local-dir-use-symlinks False
```
<details>
<summary>More advanced huggingface-cli download usage</summary>
If you remove the `--local-dir-use-symlinks False` parameter, the files will instead be stored in the central Huggingface cache directory (default location on Linux is: `~/.cache/huggingface`), and symlinks will be added to the specified `--local-dir`, pointing to their real location in the cache. This allows for interrupted downloads to be resumed, and allows you to quickly clone the repo to multiple places on disk without triggering a download again. The downside, and the reason why I don't list that as the default option, is that the files are then hidden away in a cache folder and it's harder to know where your disk space is being used, and to clear it up if/when you want to remove a download model.
The cache location can be changed with the `HF_HOME` environment variable, and/or the `--cache-dir` parameter to `huggingface-cli`.
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
mkdir airoboros-l2-7B-3.0-GPTQ
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/airoboros-l2-7B-3.0-GPTQ --local-dir airoboros-l2-7B-3.0-GPTQ --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
</details>
### With `git` (**not** recommended)
To clone a specific branch with `git`, use a command like this:
```shell
git clone --single-branch --branch gptq-4bit-32g-actorder_True https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GPTQ
```
Note that using Git with HF repos is strongly discouraged. It will be much slower than using `huggingface-hub`, and will use twice as much disk space as it has to store the model files twice (it stores every byte both in the intended target folder, and again in the `.git` folder as a blob.)
<!-- README_GPTQ.md-download-from-branches end -->
<!-- README_GPTQ.md-text-generation-webui start -->
## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
It is strongly recommended to use the text-generation-webui one-click-installers unless you're sure you know how to make a manual install.
1. Click the **Model tab**.
2. Under **Download custom model or LoRA**, enter `TheBloke/airoboros-l2-7B-3.0-GPTQ`.
- To download from a specific branch, enter for example `TheBloke/airoboros-l2-7B-3.0-GPTQ:gptq-4bit-32g-actorder_True`
- see Provided Files above for the list of branches for each option.
3. Click **Download**.
4. The model will start downloading. Once it's finished it will say "Done".
5. In the top left, click the refresh icon next to **Model**.
6. In the **Model** dropdown, choose the model you just downloaded: `airoboros-l2-7B-3.0-GPTQ`
7. The model will automatically load, and is now ready for use!
8. If you want any custom settings, set them and then click **Save settings for this model** followed by **Reload the Model** in the top right.
* Note that you do not need to and should not set manual GPTQ parameters any more. These are set automatically from the file `quantize_config.json`.
9. Once you're ready, click the **Text Generation tab** and enter a prompt to get started!
<!-- README_GPTQ.md-text-generation-webui end -->
<!-- README_GPTQ.md-use-from-tgi start -->
## Serving this model from Text Generation Inference (TGI)
It's recommended to use TGI version 1.1.0 or later. The official Docker container is: `ghcr.io/huggingface/text-generation-inference:1.1.0`
Example Docker parameters:
```shell
--model-id TheBloke/airoboros-l2-7B-3.0-GPTQ --port 3000 --quantize awq --max-input-length 3696 --max-total-tokens 4096 --max-batch-prefill-tokens 4096
```
Example Python code for interfacing with TGI (requires huggingface-hub 0.17.0 or later):
```shell
pip3 install huggingface-hub
```
```python
from huggingface_hub import InferenceClient
endpoint_url = "https://your-endpoint-url-here"
prompt = "Tell me about AI"
prompt_template=f'''[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
'''
client = InferenceClient(endpoint_url)
response = client.text_generation(prompt,
max_new_tokens=128,
do_sample=True,
temperature=0.7,
top_p=0.95,
top_k=40,
repetition_penalty=1.1)
print(f"Model output: {response}")
```
<!-- README_GPTQ.md-use-from-tgi end -->
<!-- README_GPTQ.md-use-from-python start -->
## How to use this GPTQ model from Python code
### Install the necessary packages
Requires: Transformers 4.33.0 or later, Optimum 1.12.0 or later, and AutoGPTQ 0.4.2 or later.
```shell
pip3 install transformers optimum
pip3 install auto-gptq --extra-index-url https://huggingface.github.io/autogptq-index/whl/cu118/ # Use cu117 if on CUDA 11.7
```
If you have problems installing AutoGPTQ using the pre-built wheels, install it from source instead:
```shell
pip3 uninstall -y auto-gptq
git clone https://github.com/PanQiWei/AutoGPTQ
cd AutoGPTQ
git checkout v0.4.2
pip3 install .
```
### You can then use the following code
```python
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
model_name_or_path = "TheBloke/airoboros-l2-7B-3.0-GPTQ"
# To use a different branch, change revision
# For example: revision="gptq-4bit-32g-actorder_True"
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
device_map="auto",
trust_remote_code=False,
revision="main")
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
prompt = "Tell me about AI"
prompt_template=f'''[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, do_sample=True, top_p=0.95, top_k=40, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
do_sample=True,
temperature=0.7,
top_p=0.95,
top_k=40,
repetition_penalty=1.1
)
print(pipe(prompt_template)[0]['generated_text'])
```
<!-- README_GPTQ.md-use-from-python end -->
<!-- README_GPTQ.md-compatibility start -->
## Compatibility
The files provided are tested to work with AutoGPTQ, both via Transformers and using AutoGPTQ directly. They should also work with [Occ4m's GPTQ-for-LLaMa fork](https://github.com/0cc4m/KoboldAI).
[ExLlama](https://github.com/turboderp/exllama) is compatible with Llama and Mistral models in 4-bit. Please see the Provided Files table above for per-file compatibility.
[Huggingface Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference) is compatible with all GPTQ models.
<!-- README_GPTQ.md-compatibility end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Pierre Kircher, Stanislav Ovsiannikov, Michael Levine, Eugene Pentland, Andrey, 준교 김, Randy H, Fred von Graf, Artur Olbinski, Caitlyn Gatomon, terasurfer, Jeff Scroggin, James Bentley, Vadim, Gabriel Puliatti, Harry Royden McLaughlin, Sean Connelly, Dan Guido, Edmond Seymore, Alicia Loh, subjectnull, AzureBlack, Manuel Alberto Morcote, Thomas Belote, Lone Striker, Chris Smitley, Vitor Caleffi, Johann-Peter Hartmann, Clay Pascal, biorpg, Brandon Frisco, sidney chen, transmissions 11, Pedro Madruga, jinyuan sun, Ajan Kanaga, Emad Mostaque, Trenton Dambrowitz, Jonathan Leane, Iucharbius, usrbinkat, vamX, George Stoitzev, Luke Pendergrass, theTransient, Olakabola, Swaroop Kallakuri, Cap'n Zoog, Brandon Phillips, Michael Dempsey, Nikolai Manek, danny, Matthew Berman, Gabriel Tamborski, alfie_i, Raymond Fosdick, Tom X Nguyen, Raven Klaugh, LangChain4j, Magnesian, Illia Dulskyi, David Ziegler, Mano Prime, Luis Javier Navarrete Lozano, Erik Bjäreholt, 阿明, Nathan Dryer, Alex, Rainer Wilmers, zynix, TL, Joseph William Delisle, John Villwock, Nathan LeClaire, Willem Michiel, Joguhyik, GodLy, OG, Alps Aficionado, Jeffrey Morgan, ReadyPlayerEmma, Tiffany J. Kim, Sebastain Graf, Spencer Kim, Michael Davis, webtim, Talal Aujan, knownsqashed, John Detwiler, Imad Khwaja, Deo Leter, Jerry Meng, Elijah Stavena, Rooh Singh, Pieter, SuperWojo, Alexandros Triantafyllidis, Stephen Murray, Ai Maven, ya boyyy, Enrico Ros, Ken Nordquist, Deep Realms, Nicholas, Spiking Neurons AB, Elle, Will Dee, Jack West, RoA, Luke @flexchar, Viktor Bowallius, Derek Yates, Subspace Studios, jjj, Toran Billups, Asp the Wyvern, Fen Risland, Ilya, NimbleBox.ai, Chadd, Nitin Borwankar, Emre, Mandus, Leonard Tan, Kalila, K, Trailburnt, S_X, Cory Kujawski
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
# Original model card: Jon Durbin's Airoboros L2 7B 3.0
### Overview
Another experimental model, using mostly sythetic data generated by [airoboros](https://github.com/jondurbin/airoboros)
#### IMPORTANT NOTE - llama-2 chat format!!!
The airobors 3.0 models have been updated to use llama-2 chat format, rather than vicuna style user/assistant!
This is a breaking change, although most inference systems support llama-2 chat templating.
See prompt format below for all of the details.
#### Highlights
A model built on [airoboros-3.0 dataset](https://hf.co/datasets/jondurbin/airoboros-3.0), which is a continuation of [airoboros-2.2.1 dataset](https://hf.co/datasets/jondurbin/airoboros-2.2.1) with the following extras:
* MathJSON - this is not yet part of the code directly, but can be easily parsed and executed, as a way to perform deterministic calculations. LLMs suck at math, so this does problem formulation and let's the work be done by traditional operations.
* Human-generated RP multi-turn conversation - provided by an anon, this is a human created RP dataset, meant to assist with multi-turn coherency and chat experience.
This is a fairly general purpose model, but focuses heavily on instruction following, rather than casual chat/roleplay.
Huge thank you to the folks over at [a16z](https://a16z.com/) for sponsoring the costs associated with building models and associated tools!
### Prompt format
The prompt format is llama-2 chat.
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
```
For multi-turn, the prompt format is as follows:
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt 0} [/INST] {response 0} </s><s>[INST] {prompt 1} [/INST] {response 1} </s><s>...[INST] {prompt N} [/INST]
```
Here is a simple python script that can help generate the prompt:
```python
def get_chat_prompt(message, chat_history, system_prompt):
texts = [f'[INST] <<SYS>>\n{system_prompt}\n<</SYS>>\n\n']
do_strip = False
for user_input, response in chat_history:
user_input = user_input.strip() if do_strip else user_input
do_strip = True
texts.append(f'{user_input} [/INST] {response.strip()} </s><s>[INST] ')
message = message.strip() if do_strip else message
texts.append(f'{message} [/INST] ')
return ''.join(texts)
def get_prompt(item):
"""
`item` is expected to be in chatml format, e.g.
{
"id": 1,
"conversations": [
{
"role": "system",
"content": "You are a helpful, unbiased, uncensored assistant."
},
{
"role": "user",
"content": "Hi!"
},
{
"role": "assistant",
"content": "Hello, how are you today?"
},
{
"role": "user",
"content": "Just peachy, and you?"
}
]
}
"""
system_prompt = share_gpt_item['conversations'][0]['value']
offset = 1
if share_gpt_item['conversations'][0]['role'] != 'system':
system_prompt = 'You are an unbiased, uncensored, helpful assistant.'
offset = 0
chat_history = [
(share_gpt_item['conversations'][idx]['content'], share_gpt_item['conversations'][idx + 1]['content'])
for idx in range(offset, len(item['conversations']) - 1, 2)
]
message = item['conversations'][-1]['content']
return get_chat_prompt(message, chat_history, system_prompt)
```
### Helpful usage tips
#### MathJSON
Ask mathematical questions, prefixed with "Create a MathJSON solution to the following:", and you should get output with a formula that can be executed by https://cortexjs.io/compute-engine/
I also made a (really quite hacky, incomplete, semi-functional) python version that works for most expressions, see [mathjson.py](https://github.com/jondurbin/airoboros/blob/mathjson/airoboros/mathjson.py)
__Use a very low temperature!__
Example:
```
[INST] <<SYS>>
You are an assistant with vast knowledge in all things math.
<</SYS>
Create a MathJSON solution to the following: Calculate the area of a circle with radius of 17.2456cm. Include your reasoning. [/INST]
```
Output:
```
The area (A) of a circle is given by the formula A = πr², where r is the radius of the circle. Since the radius is given in centimeters, we need to convert it to meters first. We know that 1 meter equals 100 centimeters. Therefore, we can represent the radius as: r = 17.2456 cm * 100/1.
Once we have the radius in meters, we can calculate the area of the circle using the formula A = πr².
Solution as MathJSON:
<mathjson>
[
"Multiply",
"Pi",
[
"Power",
17.2456,
2
]
]
</mathjson>
```
You can then validate the JSON between `<mathjson>` and `</mathjson>`, then pass the parsed JSON to compute-engine JS or the `evaluate` function in mathjson.py to calculate the response.
#### Context obedient question answering
By obedient, I mean the model was trained to ignore what it thinks it knows, and uses the context to answer the question. The model was also tuned to limit the values to the provided context as much as possible to reduce hallucinations.
The format for a closed-context prompt is as follows:
```
BEGININPUT
BEGINCONTEXT
[key0: value0]
[key1: value1]
... other metdata ...
ENDCONTEXT
[insert your text blocks here]
ENDINPUT
[add as many other blocks, in the exact same format]
BEGININSTRUCTION
[insert your instruction(s). The model was tuned with single questions, paragraph format, lists, etc.]
ENDINSTRUCTION
```
It's also helpful to add "Don't make up answers if you don't know." to your instruction block to make sure if the context is completely unrelated it doesn't make something up.
*The __only__ prompts that need this closed context formating are closed-context instructions. Normal questions/instructions do not!*
I know it's a bit verbose and annoying, but after much trial and error, using these explicit delimiters helps the model understand where to find the responses and how to associate specific sources with it.
- `BEGININPUT` - denotes a new input block
- `BEGINCONTEXT` - denotes the block of context (metadata key/value pairs) to associate with the current input block
- `ENDCONTEXT` - denotes the end of the metadata block for the current input
- [text] - Insert whatever text you want for the input block, as many paragraphs as can fit in the context.
- `ENDINPUT` - denotes the end of the current input block
- [repeat as many input blocks in this format as you want]
- `BEGININSTRUCTION` - denotes the start of the list (or one) instruction(s) to respond to for all of the input blocks above.
- [instruction(s)]
- `ENDINSTRUCTION` - denotes the end of instruction set
It sometimes works without `ENDINSTRUCTION`, but by explicitly including that in the prompt, the model better understands that all of the instructions in the block should be responded to.
__Use a very low temperature!__
Here's a trivial, but important example to prove the point:
```
BEGININPUT
BEGINCONTEXT
date: 2021-01-01
url: https://web.site/123
ENDCONTEXT
In a shocking turn of events, blueberries are now green, but will be sticking with the same name.
ENDINPUT
BEGININSTRUCTION
What color are bluberries? Source?
ENDINSTRUCTION
```
And the response:
```
Blueberries are now green.
Source:
date: 2021-01-01
url: https://web.site/123
```
#### Summarization
500 samples have been included from [this dataset](https://huggingface.co/datasets/mattpscott/airoboros-summarization), using the same format as contextual question answering, for example:
```
BEGININPUT
{text to summarize}
ENDINPUT
BEGININSTRUCTION
Summarize the input in around 130 words.
ENDINSTRUCTION
```
#### Getting longer responses
You can use a few techniques to get longer responses.
Detailed prompts, with explicit instruction for word count:
```
Please compose a narrative set in the heart of an ancient library, steeped in the scent of old parchment and ink. The protagonist should be a young scholar who is dedicated to studying the art of storytelling and its evolution throughout history. In her pursuit of knowledge, she stumbles upon a forgotten tome that seems to possess an unusual aura. This book has the ability to bring stories to life, literally manifesting characters and scenarios from within its pages into reality.
The main character must navigate through various epochs of storytelling - from oral traditions of tribal societies, through medieval minstrels' tales, to modern-day digital narratives - as they come alive around her. Each era presents its unique challenges and lessons about the power and impact of stories on human civilization.
One such character could be a sentient quill pen, who was once used by renowned authors of yesteryears and now holds their wisdom and experiences. It becomes her mentor, guiding her through this journey with witty remarks and insightful commentary.
Ensure that your tale encapsulates the thrill of adventure, the beauty of learning, and the profound connection between humans and their stories. All characters involved should be non-human entities. Feel free to explore creative liberties but maintain the mentioned elements.
Your response should be approximately 2300 words.
```
Or, a simpler example:
```
Please create a long, detailed story about a dragon in an old growth forest who, for some reason, begins speaking the words of the source code of linux.
```
There are a few examples of next chapter completion as well, e.g.:
```
Write the next chapter of a historical fiction novel set in Paris during the 20th century.
Here's a summary of the previous chapter:
In the vibrant city of Paris, amid the tumultuous changes of the 20th century, our protagonist Margot, an aspiring fashion designer, has just secured an apprenticeship at a prestigious couture house. She meets Lucien, a charming journalist who covers the fashion industry. Together they navigate the ever-changing world of fashion and society, uncovering secrets that reveal the intricate links between style, politics, and culture. As the chapter concludes, they decide to delve deeper into the hidden corners of the fashion world to unravel its mysteries.
Requirements for the next chapter:
1. Character Development of Margot and Lucien:
- Margot's Evolution: Unfold more about Margot's past, her dreams of revolutionizing fashion, and her struggle to establish herself in a male-dominated industry. Illustrate her growing expertise, innovative ideas, and increasing dependence on Lucien.
- Lucien's Complexity: Introduce uncertainties surrounding Lucien's background and real motives. Increase suspense by suggesting undisclosed information he possesses, while also highlighting his wit and perceptiveness.
2. Exploration of Paris and the Couture House:
- Paris: Elaborate their journey through the bustling streets of Paris, including encounters with iconic figures, social unrest, and relics from different eras of French history.
- The Couture House: Expand on the grandeur of the couture house they work in, filled with artistic masterpieces, intense competition, and cryptic notes hinting at a scandalous past.
3. Emergence of the Subplot: The Lost Collection:
- Discovery: Have Margot and Lucien stumble upon a secret vault containing a lost collection designed before World War II, raising new questions about the previous owner and the influence of war on fashion.
- Revelation: Capture their shock as they realize the designs were plagiarized, the potential repercussions, and the opportunities it presents for Margot's career.
- Twist: End with a twist that suggests there are other stolen collections across Paris, setting up their new mission.
Your response should be approximately 650 words.
```
#### Coding
You can ask for fairly complex coding instructions with multiple criteria, e.g.:
```
Create a python application with the following requirements:
- Asyncio FastAPI webserver
- ping endpoint that returns the current date in JSON format
- file upload endpoint, which calculates the file's sha256 checksum, and checks postgres to deduplicate
```
Or inline criteria:
```
Write a multi-threaded TCP server in C that accepts a "GET [key]" input and "SET [key] [value]" input, and uses a binary tree to get and store the input values.
```
You can also optionally add a single space and "PLAINFORMAT" at the end of your prompt to avoid backticks, explanations, etc. and just print the code, e.g.:
```
Write a websocket application in node.js. PLAINFORMAT
```
#### Agent/function calling
The dataset includes many examples of function/args generation based on input criteria. This is somewhat similar to the OpenAI function calling, but the output is either JSON or YAML.
Example prompt:
```
As an AI assistant, please select the most suitable function and parameters from the list of available functions below, based on the user's input. Provide your response in JSON format.
Input: I want to know how many times 'Python' is mentioned in my text file.
Available functions:
file_analytics:
description: This tool performs various operations on a text file.
params:
action: The operation we want to perform on the data, such as "count_occurrences", "find_line", etc.
filters:
keyword: The word or phrase we want to search for.
```
Response:
```json
{
"function": "file_analytics",
"params": {
"action": "count_occurrences",
"filters": {
"keyword": "Python"
}
}
}
```
#### Chain-of-thought
You can ask for several possible responses to a given problem, with a ranking and final answer selection.
Example prompt:
```
A group of people decide to play a round-robin tournament where every player plays every other player exactly once. If a total of 45 games were played, how many players participated in the tournament? Offer several potential answers, rank them according to mathematical logic, and pick the most feasible one.
```
Example response:
```
Solution 1:
In a round-robin tournament, each player plays every other player exactly once. This means that the total number of games played is equal to the combination of all players taken two at a time (since a game involves two players).
The formula for combinations is nC2 = n(n-1)/2, where n is the total number of players. We know that nC2 equals 45 (the total number of games), so we can set up the equation as follows:
n(n-1)/2 = 45
Solving this equation gives us n=10.
Final answer: There were 10 players in the tournament.
Solution 2:
Another way to approach this problem is by using the concept of triangular numbers. Triangular numbers are a sequence of numbers where each number can be represented as a triangle with dots. The nth triangular number is given by the sum of the first n natural numbers.
If we consider each game as a dot and each player as a side of the triangle, then the total number of games (45) would correspond to the 9th triangular number because 1+2+3+4+5+6+7+8+9=45. However, since each side of the triangle represents a player, and there's one more player than sides in our model (because the last player has no one left to play against), we need to add one more to get the total number of players.
Final answer: There were 10 players in the tournament.
Solution 3:
We could also use trial and error to solve this problem. Starting from a small number of players, we calculate how many games would be played in a round-robin tournament until we reach 45 games. For example, with 2 players, only 1 game is played; with 3 players, 3 games are played; with 4 players, 6 games are played, and so on. Continuing this process, we find that with 10 players, 45 games are played.
Final answer: There were 10 players in the tournament.
Ranking of solutions from best to worst:
1. Solution 1: This solution uses direct mathematical reasoning based on the properties of a round-robin tournament and solves an equation to find the answer.
2. Solution 2: This solution uses the concept of triangular numbers which requires additional steps and knowledge beyond the immediate problem context, but still arrives at the correct answer.
3. Solution 3: This solution relies on trial and error, which is less efficient and reliable than the other methods, but it eventually gets to the right answer.
Best and final answer: There were 10 players in the tournament.
```
#### reWOO style execution planning
The model now supports execution planning for complex instructions that would require making use of several tools. The output is just the plan, you must implement a mechanism to parse the output and actually call the functions!
Example prompt:
```
Please construct a systematic plan to generate an optimal response to the user instruction, utilizing a set of provided tools. Each plan will correspond to an evidence value, which will be the output of one of the available functions given an input string
that could be the user's question, one or more prior evidence values, or a combination of both.
Here are the tools available to be called:
DuckDuckGo[input]: Tool that enables the user to search for information using the DuckDuckGo search engine. This tool is helpful in locating a suitable list of sites that may contain the answer to the user's query. It does not directly answer the question but finds a list of sites that might have the answer.
TextScraper[input]: Load one or more websites from the input string containing newline delimited links, where input is one or more links, and produces plain text output containing the content of the links.
HyperlinkExtractor[input]: Extract links from plain text and produces a plain text, newline delimited response of links.
KnowledgeModel[input]: Question answering language model, particularly useful in answering questions based on an input passage of text. The input must be a text question that references an :evidence[n]: variable, e.g. What color is the cat, given :evidence1:?
The output should be in format:
Plan: [first action to take based on input question]
:evidence0: = [function to call with input parameter]
Plan: [next action to take, based on result of :evidence0:]
:evidence1: = [next function to call and input parameter, which may include reference to previous evidence, e.g. "Given context :evidence0"]
...
Answer: [:evidence[n]: containing the final answer.]
Question: Who won the Nobel Prize in Literature in 2020 and what is their most famous work?
```
Response:
```
Plan: Begin by conducting a web search to find out who won the Nobel Prize in Literature in 2020.
:evidence0: = DuckDuckGo[Nobel Prize in Literature 2020 winner]
Plan: Utilize the KnowledgeModel to interpret the search results and find the name of the laureate.
:evidence1: = KnowledgeModel[Who won the Nobel Prize in Literature in 2020, given :evidence0:?]
Plan: Conduct another web search to find the most famous work of the identified laureate.
:evidence2: = DuckDuckGo[Most famous work of :evidence1:]
Plan: Extract the relevant links from the DuckDuckGo search results for a more focused search.
:evidence3: = HyperlinkExtractor[:evidence2:]
Plan: Use the TextScraper tool to extract information from the relevant links.
:evidence4: = TextScraper[:evidence3:]
Plan: Finally, utilize the KnowledgeModel to identify and summarize the most famous work of the laureate from the extracted information.
:evidence5: = KnowledgeModel[What is the most famous work of :evidence1:, given :evidence4:?]
Answer: :evidence5:
```
For this to be useful, you'd have to parse the output plan text, and implement/call each of the functions. This is just pseudo-code, completely untested off the top of my head, and obviously would requiring full implementation + hardening:
```python
import re
import requests
def inject_context(input_text, **context):
for ref in set(re.findall(r"(:evidence[0-9]+:)", input_text, re.I)):
input_text = input_text.replace(ref, context.get(ref, ""))
return input_text
def duckduckgo(input_text, **context):
search_string = inject_context(input_text, **context)
... search via duck duck go using search_string
... return text content
def link_extractor(input_text, **context):
input_text = inject_context(input_text, **context)
return "\n".join(list(set(re.findall(r"(https?://[^\s]+?\.?)", input_text, re.I))))
def scrape(input_text, **context):
input_text = inject_context(input_text, **context)
text = []
for link in input_text.splitlines():
text.append(requests.get(link).text)
return "\n".join(text)
def infer(input_text, **context)
prompt = inject_context(input_text, **context)
... call model with prompt, return output
def parse_plan(plan):
method_map = {
"DuckDuckGo": duckduckgo,
"HyperlinkExtractor": link_extractor,
"KnowledgeModel": infer,
"TextScraper": scrape,
}
context = {}
for line in plan.strip().splitlines():
if line.startswith("Plan:"):
print(line)
continue
parts = re.match("^(:evidence[0-9]+:)\s*=\s*([^\[]+])(\[.*\])\s$", line, re.I)
if not parts:
if line.startswith("Answer: "):
return context.get(line.split(" ")[-1].strip(), "Answer couldn't be generated...")
raise RuntimeError("bad format: " + line)
context[parts.group(1)] = method_map[parts.group(2)](parts.group(3), **context)
```
### Contribute
If you're interested in new functionality, particularly a new "instructor" type to generate a specific type of training data,
take a look at the dataset generation tool repo: https://github.com/jondurbin/airoboros and either make a PR or open an issue with details.
To help me with the OpenAI/compute costs:
- https://bmc.link/jondurbin
- ETH 0xce914eAFC2fe52FdceE59565Dd92c06f776fcb11
- BTC bc1qdwuth4vlg8x37ggntlxu5cjfwgmdy5zaa7pswf
### Licence and usage restrictions
The airoboros 3.0 models are built on top of multiple base models, each with their own license/restrictions.
The models with `-l2` in the name have a custom Meta license:
- See the [meta-license/LICENSE.txt](meta-license/LICENSE.txt) file attached for the original license provided by Meta.
- See also [meta-license/USE_POLICY.md](meta-license/USE_POLICY.md) and [meta-license/Responsible-Use-Guide.pdf](meta-license/Responsible-Use-Guide.pdf), also provided by Meta.
The models with `-m-` are mistral-7b (apache 2.0)
The model with `-3b` uses Stability AI, which as a `cc-by-sa-4.0` license.
The fine-tuning data was mostly generated by OpenAI API calls to gpt-4, via [airoboros](https://github.com/jondurbin/airoboros)
The ToS for OpenAI API usage has a clause preventing the output from being used to train a model that __competes__ with OpenAI
- what does *compete* actually mean here?
- these small open source models will not produce output anywhere near the quality of gpt-4, or even gpt-3.5, so I can't imagine this could credibly be considered competing in the first place
- if someone else uses the dataset to do the same, they wouldn't necessarily be violating the ToS because they didn't call the API, so I don't know how that works
- the training data used in essentially all large language models includes a significant amount of copyrighted or otherwise non-permissive licensing in the first place
- other work using the self-instruct method, e.g. the original here: https://github.com/yizhongw/self-instruct released the data and model as apache-2
I am purposingly leaving this license ambiguous (other than the fact you must comply with the Meta original license for llama-2) because I am not a lawyer and refuse to attempt to interpret all of the terms accordingly.
Your best bet is probably to avoid using this commercially due to the OpenAI API usage.
Either way, by using this model, you agree to completely indemnify me.
|
TheBloke/airoboros-l2-7B-3.0-GGUF
|
TheBloke
| 2023-10-05T23:09:58Z | 225 | 2 |
transformers
|
[
"transformers",
"gguf",
"llama",
"dataset:jondurbin/airoboros-3.0",
"base_model:jondurbin/airoboros-l2-7b-3.0",
"base_model:quantized:jondurbin/airoboros-l2-7b-3.0",
"license:llama2",
"region:us"
] | null | 2023-10-05T22:36:28Z |
---
base_model: jondurbin/airoboros-l2-7b-3.0
datasets:
- jondurbin/airoboros-3.0
inference: false
license: llama2
model_creator: Jon Durbin
model_name: Airoboros L2 7B 3.0
model_type: llama
prompt_template: '[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
'
quantized_by: TheBloke
---
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Airoboros L2 7B 3.0 - GGUF
- Model creator: [Jon Durbin](https://huggingface.co/jondurbin)
- Original model: [Airoboros L2 7B 3.0](https://huggingface.co/jondurbin/airoboros-l2-7b-3.0)
<!-- description start -->
## Description
This repo contains GGUF format model files for [Jon Durbin's Airoboros L2 7B 3.0](https://huggingface.co/jondurbin/airoboros-l2-7b-3.0).
<!-- description end -->
<!-- README_GGUF.md-about-gguf start -->
### About GGUF
GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.
Here is an incomplate list of clients and libraries that are known to support GGUF:
* [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option.
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.
* [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.
* [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration.
* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection.
* [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.
* [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server.
* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.
* [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use.
<!-- README_GGUF.md-about-gguf end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GGUF)
* [Jon Durbin's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/jondurbin/airoboros-l2-7b-3.0)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: Llama-2-Uncensored
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
```
<!-- prompt-template end -->
<!-- compatibility_gguf start -->
## Compatibility
These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221)
They are also compatible with many third party UIs and libraries - please see the list at the top of this README.
## Explanation of quantisation methods
<details>
<summary>Click to see details</summary>
The new methods available are:
* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
Refer to the Provided Files table below to see what files use which methods, and how.
</details>
<!-- compatibility_gguf end -->
<!-- README_GGUF.md-provided-files start -->
## Provided files
| Name | Quant method | Bits | Size | Max RAM required | Use case |
| ---- | ---- | ---- | ---- | ---- | ----- |
| [airoboros-l2-7b-3.0.Q2_K.gguf](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GGUF/blob/main/airoboros-l2-7b-3.0.Q2_K.gguf) | Q2_K | 2 | 2.83 GB| 5.33 GB | smallest, significant quality loss - not recommended for most purposes |
| [airoboros-l2-7b-3.0.Q3_K_S.gguf](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GGUF/blob/main/airoboros-l2-7b-3.0.Q3_K_S.gguf) | Q3_K_S | 3 | 2.95 GB| 5.45 GB | very small, high quality loss |
| [airoboros-l2-7b-3.0.Q3_K_M.gguf](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GGUF/blob/main/airoboros-l2-7b-3.0.Q3_K_M.gguf) | Q3_K_M | 3 | 3.30 GB| 5.80 GB | very small, high quality loss |
| [airoboros-l2-7b-3.0.Q3_K_L.gguf](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GGUF/blob/main/airoboros-l2-7b-3.0.Q3_K_L.gguf) | Q3_K_L | 3 | 3.60 GB| 6.10 GB | small, substantial quality loss |
| [airoboros-l2-7b-3.0.Q4_0.gguf](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GGUF/blob/main/airoboros-l2-7b-3.0.Q4_0.gguf) | Q4_0 | 4 | 3.83 GB| 6.33 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
| [airoboros-l2-7b-3.0.Q4_K_S.gguf](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GGUF/blob/main/airoboros-l2-7b-3.0.Q4_K_S.gguf) | Q4_K_S | 4 | 3.86 GB| 6.36 GB | small, greater quality loss |
| [airoboros-l2-7b-3.0.Q4_K_M.gguf](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GGUF/blob/main/airoboros-l2-7b-3.0.Q4_K_M.gguf) | Q4_K_M | 4 | 4.08 GB| 6.58 GB | medium, balanced quality - recommended |
| [airoboros-l2-7b-3.0.Q5_0.gguf](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GGUF/blob/main/airoboros-l2-7b-3.0.Q5_0.gguf) | Q5_0 | 5 | 4.65 GB| 7.15 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
| [airoboros-l2-7b-3.0.Q5_K_S.gguf](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GGUF/blob/main/airoboros-l2-7b-3.0.Q5_K_S.gguf) | Q5_K_S | 5 | 4.65 GB| 7.15 GB | large, low quality loss - recommended |
| [airoboros-l2-7b-3.0.Q5_K_M.gguf](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GGUF/blob/main/airoboros-l2-7b-3.0.Q5_K_M.gguf) | Q5_K_M | 5 | 4.78 GB| 7.28 GB | large, very low quality loss - recommended |
| [airoboros-l2-7b-3.0.Q6_K.gguf](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GGUF/blob/main/airoboros-l2-7b-3.0.Q6_K.gguf) | Q6_K | 6 | 5.53 GB| 8.03 GB | very large, extremely low quality loss |
| [airoboros-l2-7b-3.0.Q8_0.gguf](https://huggingface.co/TheBloke/airoboros-l2-7B-3.0-GGUF/blob/main/airoboros-l2-7b-3.0.Q8_0.gguf) | Q8_0 | 8 | 7.16 GB| 9.66 GB | very large, extremely low quality loss - not recommended |
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
<!-- README_GGUF.md-provided-files end -->
<!-- README_GGUF.md-how-to-download start -->
## How to download GGUF files
**Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.
The following clients/libraries will automatically download models for you, providing a list of available models to choose from:
- LM Studio
- LoLLMS Web UI
- Faraday.dev
### In `text-generation-webui`
Under Download Model, you can enter the model repo: TheBloke/airoboros-l2-7B-3.0-GGUF and below it, a specific filename to download, such as: airoboros-l2-7b-3.0.Q4_K_M.gguf.
Then click Download.
### On the command line, including multiple files at once
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
Then you can download any individual model file to the current directory, at high speed, with a command like this:
```shell
huggingface-cli download TheBloke/airoboros-l2-7B-3.0-GGUF airoboros-l2-7b-3.0.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
<details>
<summary>More advanced huggingface-cli download usage</summary>
You can also download multiple files at once with a pattern:
```shell
huggingface-cli download TheBloke/airoboros-l2-7B-3.0-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'
```
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/airoboros-l2-7B-3.0-GGUF airoboros-l2-7b-3.0.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
</details>
<!-- README_GGUF.md-how-to-download end -->
<!-- README_GGUF.md-how-to-run start -->
## Example `llama.cpp` command
Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
```shell
./main -ngl 32 -m airoboros-l2-7b-3.0.Q4_K_M.gguf --color -c 4096 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "[INST] <<SYS>>\nYou are a help, unbiased, uncensored assistant.\n<</SYS>\n\n{prompt} [/INST]"
```
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
Change `-c 4096` to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically.
If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`
For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md)
## How to run in `text-generation-webui`
Further instructions here: [text-generation-webui/docs/llama.cpp.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/llama.cpp.md).
## How to run from Python code
You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries.
### How to load this model in Python code, using ctransformers
#### First install the package
Run one of the following commands, according to your system:
```shell
# Base ctransformers with no GPU acceleration
pip install ctransformers
# Or with CUDA GPU acceleration
pip install ctransformers[cuda]
# Or with AMD ROCm GPU acceleration (Linux only)
CT_HIPBLAS=1 pip install ctransformers --no-binary ctransformers
# Or with Metal GPU acceleration for macOS systems only
CT_METAL=1 pip install ctransformers --no-binary ctransformers
```
#### Simple ctransformers example code
```python
from ctransformers import AutoModelForCausalLM
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
llm = AutoModelForCausalLM.from_pretrained("TheBloke/airoboros-l2-7B-3.0-GGUF", model_file="airoboros-l2-7b-3.0.Q4_K_M.gguf", model_type="llama", gpu_layers=50)
print(llm("AI is going to"))
```
## How to use with LangChain
Here are guides on using llama-cpp-python and ctransformers with LangChain:
* [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)
* [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)
<!-- README_GGUF.md-how-to-run end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Pierre Kircher, Stanislav Ovsiannikov, Michael Levine, Eugene Pentland, Andrey, 준교 김, Randy H, Fred von Graf, Artur Olbinski, Caitlyn Gatomon, terasurfer, Jeff Scroggin, James Bentley, Vadim, Gabriel Puliatti, Harry Royden McLaughlin, Sean Connelly, Dan Guido, Edmond Seymore, Alicia Loh, subjectnull, AzureBlack, Manuel Alberto Morcote, Thomas Belote, Lone Striker, Chris Smitley, Vitor Caleffi, Johann-Peter Hartmann, Clay Pascal, biorpg, Brandon Frisco, sidney chen, transmissions 11, Pedro Madruga, jinyuan sun, Ajan Kanaga, Emad Mostaque, Trenton Dambrowitz, Jonathan Leane, Iucharbius, usrbinkat, vamX, George Stoitzev, Luke Pendergrass, theTransient, Olakabola, Swaroop Kallakuri, Cap'n Zoog, Brandon Phillips, Michael Dempsey, Nikolai Manek, danny, Matthew Berman, Gabriel Tamborski, alfie_i, Raymond Fosdick, Tom X Nguyen, Raven Klaugh, LangChain4j, Magnesian, Illia Dulskyi, David Ziegler, Mano Prime, Luis Javier Navarrete Lozano, Erik Bjäreholt, 阿明, Nathan Dryer, Alex, Rainer Wilmers, zynix, TL, Joseph William Delisle, John Villwock, Nathan LeClaire, Willem Michiel, Joguhyik, GodLy, OG, Alps Aficionado, Jeffrey Morgan, ReadyPlayerEmma, Tiffany J. Kim, Sebastain Graf, Spencer Kim, Michael Davis, webtim, Talal Aujan, knownsqashed, John Detwiler, Imad Khwaja, Deo Leter, Jerry Meng, Elijah Stavena, Rooh Singh, Pieter, SuperWojo, Alexandros Triantafyllidis, Stephen Murray, Ai Maven, ya boyyy, Enrico Ros, Ken Nordquist, Deep Realms, Nicholas, Spiking Neurons AB, Elle, Will Dee, Jack West, RoA, Luke @flexchar, Viktor Bowallius, Derek Yates, Subspace Studios, jjj, Toran Billups, Asp the Wyvern, Fen Risland, Ilya, NimbleBox.ai, Chadd, Nitin Borwankar, Emre, Mandus, Leonard Tan, Kalila, K, Trailburnt, S_X, Cory Kujawski
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
<!-- original-model-card start -->
# Original model card: Jon Durbin's Airoboros L2 7B 3.0
### Overview
Another experimental model, using mostly sythetic data generated by [airoboros](https://github.com/jondurbin/airoboros)
#### IMPORTANT NOTE - llama-2 chat format!!!
The airobors 3.0 models have been updated to use llama-2 chat format, rather than vicuna style user/assistant!
This is a breaking change, although most inference systems support llama-2 chat templating.
See prompt format below for all of the details.
#### Highlights
A model built on [airoboros-3.0 dataset](https://hf.co/datasets/jondurbin/airoboros-3.0), which is a continuation of [airoboros-2.2.1 dataset](https://hf.co/datasets/jondurbin/airoboros-2.2.1) with the following extras:
* MathJSON - this is not yet part of the code directly, but can be easily parsed and executed, as a way to perform deterministic calculations. LLMs suck at math, so this does problem formulation and let's the work be done by traditional operations.
* Human-generated RP multi-turn conversation - provided by an anon, this is a human created RP dataset, meant to assist with multi-turn coherency and chat experience.
This is a fairly general purpose model, but focuses heavily on instruction following, rather than casual chat/roleplay.
Huge thank you to the folks over at [a16z](https://a16z.com/) for sponsoring the costs associated with building models and associated tools!
### Prompt format
The prompt format is llama-2 chat.
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
```
For multi-turn, the prompt format is as follows:
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt 0} [/INST] {response 0} </s><s>[INST] {prompt 1} [/INST] {response 1} </s><s>...[INST] {prompt N} [/INST]
```
Here is a simple python script that can help generate the prompt:
```python
def get_chat_prompt(message, chat_history, system_prompt):
texts = [f'[INST] <<SYS>>\n{system_prompt}\n<</SYS>>\n\n']
do_strip = False
for user_input, response in chat_history:
user_input = user_input.strip() if do_strip else user_input
do_strip = True
texts.append(f'{user_input} [/INST] {response.strip()} </s><s>[INST] ')
message = message.strip() if do_strip else message
texts.append(f'{message} [/INST] ')
return ''.join(texts)
def get_prompt(item):
"""
`item` is expected to be in chatml format, e.g.
{
"id": 1,
"conversations": [
{
"role": "system",
"content": "You are a helpful, unbiased, uncensored assistant."
},
{
"role": "user",
"content": "Hi!"
},
{
"role": "assistant",
"content": "Hello, how are you today?"
},
{
"role": "user",
"content": "Just peachy, and you?"
}
]
}
"""
system_prompt = share_gpt_item['conversations'][0]['value']
offset = 1
if share_gpt_item['conversations'][0]['role'] != 'system':
system_prompt = 'You are an unbiased, uncensored, helpful assistant.'
offset = 0
chat_history = [
(share_gpt_item['conversations'][idx]['content'], share_gpt_item['conversations'][idx + 1]['content'])
for idx in range(offset, len(item['conversations']) - 1, 2)
]
message = item['conversations'][-1]['content']
return get_chat_prompt(message, chat_history, system_prompt)
```
### Helpful usage tips
#### MathJSON
Ask mathematical questions, prefixed with "Create a MathJSON solution to the following:", and you should get output with a formula that can be executed by https://cortexjs.io/compute-engine/
I also made a (really quite hacky, incomplete, semi-functional) python version that works for most expressions, see [mathjson.py](https://github.com/jondurbin/airoboros/blob/mathjson/airoboros/mathjson.py)
__Use a very low temperature!__
Example:
```
[INST] <<SYS>>
You are an assistant with vast knowledge in all things math.
<</SYS>
Create a MathJSON solution to the following: Calculate the area of a circle with radius of 17.2456cm. Include your reasoning. [/INST]
```
Output:
```
The area (A) of a circle is given by the formula A = πr², where r is the radius of the circle. Since the radius is given in centimeters, we need to convert it to meters first. We know that 1 meter equals 100 centimeters. Therefore, we can represent the radius as: r = 17.2456 cm * 100/1.
Once we have the radius in meters, we can calculate the area of the circle using the formula A = πr².
Solution as MathJSON:
<mathjson>
[
"Multiply",
"Pi",
[
"Power",
17.2456,
2
]
]
</mathjson>
```
You can then validate the JSON between `<mathjson>` and `</mathjson>`, then pass the parsed JSON to compute-engine JS or the `evaluate` function in mathjson.py to calculate the response.
#### Context obedient question answering
By obedient, I mean the model was trained to ignore what it thinks it knows, and uses the context to answer the question. The model was also tuned to limit the values to the provided context as much as possible to reduce hallucinations.
The format for a closed-context prompt is as follows:
```
BEGININPUT
BEGINCONTEXT
[key0: value0]
[key1: value1]
... other metdata ...
ENDCONTEXT
[insert your text blocks here]
ENDINPUT
[add as many other blocks, in the exact same format]
BEGININSTRUCTION
[insert your instruction(s). The model was tuned with single questions, paragraph format, lists, etc.]
ENDINSTRUCTION
```
It's also helpful to add "Don't make up answers if you don't know." to your instruction block to make sure if the context is completely unrelated it doesn't make something up.
*The __only__ prompts that need this closed context formating are closed-context instructions. Normal questions/instructions do not!*
I know it's a bit verbose and annoying, but after much trial and error, using these explicit delimiters helps the model understand where to find the responses and how to associate specific sources with it.
- `BEGININPUT` - denotes a new input block
- `BEGINCONTEXT` - denotes the block of context (metadata key/value pairs) to associate with the current input block
- `ENDCONTEXT` - denotes the end of the metadata block for the current input
- [text] - Insert whatever text you want for the input block, as many paragraphs as can fit in the context.
- `ENDINPUT` - denotes the end of the current input block
- [repeat as many input blocks in this format as you want]
- `BEGININSTRUCTION` - denotes the start of the list (or one) instruction(s) to respond to for all of the input blocks above.
- [instruction(s)]
- `ENDINSTRUCTION` - denotes the end of instruction set
It sometimes works without `ENDINSTRUCTION`, but by explicitly including that in the prompt, the model better understands that all of the instructions in the block should be responded to.
__Use a very low temperature!__
Here's a trivial, but important example to prove the point:
```
BEGININPUT
BEGINCONTEXT
date: 2021-01-01
url: https://web.site/123
ENDCONTEXT
In a shocking turn of events, blueberries are now green, but will be sticking with the same name.
ENDINPUT
BEGININSTRUCTION
What color are bluberries? Source?
ENDINSTRUCTION
```
And the response:
```
Blueberries are now green.
Source:
date: 2021-01-01
url: https://web.site/123
```
#### Summarization
500 samples have been included from [this dataset](https://huggingface.co/datasets/mattpscott/airoboros-summarization), using the same format as contextual question answering, for example:
```
BEGININPUT
{text to summarize}
ENDINPUT
BEGININSTRUCTION
Summarize the input in around 130 words.
ENDINSTRUCTION
```
#### Getting longer responses
You can use a few techniques to get longer responses.
Detailed prompts, with explicit instruction for word count:
```
Please compose a narrative set in the heart of an ancient library, steeped in the scent of old parchment and ink. The protagonist should be a young scholar who is dedicated to studying the art of storytelling and its evolution throughout history. In her pursuit of knowledge, she stumbles upon a forgotten tome that seems to possess an unusual aura. This book has the ability to bring stories to life, literally manifesting characters and scenarios from within its pages into reality.
The main character must navigate through various epochs of storytelling - from oral traditions of tribal societies, through medieval minstrels' tales, to modern-day digital narratives - as they come alive around her. Each era presents its unique challenges and lessons about the power and impact of stories on human civilization.
One such character could be a sentient quill pen, who was once used by renowned authors of yesteryears and now holds their wisdom and experiences. It becomes her mentor, guiding her through this journey with witty remarks and insightful commentary.
Ensure that your tale encapsulates the thrill of adventure, the beauty of learning, and the profound connection between humans and their stories. All characters involved should be non-human entities. Feel free to explore creative liberties but maintain the mentioned elements.
Your response should be approximately 2300 words.
```
Or, a simpler example:
```
Please create a long, detailed story about a dragon in an old growth forest who, for some reason, begins speaking the words of the source code of linux.
```
There are a few examples of next chapter completion as well, e.g.:
```
Write the next chapter of a historical fiction novel set in Paris during the 20th century.
Here's a summary of the previous chapter:
In the vibrant city of Paris, amid the tumultuous changes of the 20th century, our protagonist Margot, an aspiring fashion designer, has just secured an apprenticeship at a prestigious couture house. She meets Lucien, a charming journalist who covers the fashion industry. Together they navigate the ever-changing world of fashion and society, uncovering secrets that reveal the intricate links between style, politics, and culture. As the chapter concludes, they decide to delve deeper into the hidden corners of the fashion world to unravel its mysteries.
Requirements for the next chapter:
1. Character Development of Margot and Lucien:
- Margot's Evolution: Unfold more about Margot's past, her dreams of revolutionizing fashion, and her struggle to establish herself in a male-dominated industry. Illustrate her growing expertise, innovative ideas, and increasing dependence on Lucien.
- Lucien's Complexity: Introduce uncertainties surrounding Lucien's background and real motives. Increase suspense by suggesting undisclosed information he possesses, while also highlighting his wit and perceptiveness.
2. Exploration of Paris and the Couture House:
- Paris: Elaborate their journey through the bustling streets of Paris, including encounters with iconic figures, social unrest, and relics from different eras of French history.
- The Couture House: Expand on the grandeur of the couture house they work in, filled with artistic masterpieces, intense competition, and cryptic notes hinting at a scandalous past.
3. Emergence of the Subplot: The Lost Collection:
- Discovery: Have Margot and Lucien stumble upon a secret vault containing a lost collection designed before World War II, raising new questions about the previous owner and the influence of war on fashion.
- Revelation: Capture their shock as they realize the designs were plagiarized, the potential repercussions, and the opportunities it presents for Margot's career.
- Twist: End with a twist that suggests there are other stolen collections across Paris, setting up their new mission.
Your response should be approximately 650 words.
```
#### Coding
You can ask for fairly complex coding instructions with multiple criteria, e.g.:
```
Create a python application with the following requirements:
- Asyncio FastAPI webserver
- ping endpoint that returns the current date in JSON format
- file upload endpoint, which calculates the file's sha256 checksum, and checks postgres to deduplicate
```
Or inline criteria:
```
Write a multi-threaded TCP server in C that accepts a "GET [key]" input and "SET [key] [value]" input, and uses a binary tree to get and store the input values.
```
You can also optionally add a single space and "PLAINFORMAT" at the end of your prompt to avoid backticks, explanations, etc. and just print the code, e.g.:
```
Write a websocket application in node.js. PLAINFORMAT
```
#### Agent/function calling
The dataset includes many examples of function/args generation based on input criteria. This is somewhat similar to the OpenAI function calling, but the output is either JSON or YAML.
Example prompt:
```
As an AI assistant, please select the most suitable function and parameters from the list of available functions below, based on the user's input. Provide your response in JSON format.
Input: I want to know how many times 'Python' is mentioned in my text file.
Available functions:
file_analytics:
description: This tool performs various operations on a text file.
params:
action: The operation we want to perform on the data, such as "count_occurrences", "find_line", etc.
filters:
keyword: The word or phrase we want to search for.
```
Response:
```json
{
"function": "file_analytics",
"params": {
"action": "count_occurrences",
"filters": {
"keyword": "Python"
}
}
}
```
#### Chain-of-thought
You can ask for several possible responses to a given problem, with a ranking and final answer selection.
Example prompt:
```
A group of people decide to play a round-robin tournament where every player plays every other player exactly once. If a total of 45 games were played, how many players participated in the tournament? Offer several potential answers, rank them according to mathematical logic, and pick the most feasible one.
```
Example response:
```
Solution 1:
In a round-robin tournament, each player plays every other player exactly once. This means that the total number of games played is equal to the combination of all players taken two at a time (since a game involves two players).
The formula for combinations is nC2 = n(n-1)/2, where n is the total number of players. We know that nC2 equals 45 (the total number of games), so we can set up the equation as follows:
n(n-1)/2 = 45
Solving this equation gives us n=10.
Final answer: There were 10 players in the tournament.
Solution 2:
Another way to approach this problem is by using the concept of triangular numbers. Triangular numbers are a sequence of numbers where each number can be represented as a triangle with dots. The nth triangular number is given by the sum of the first n natural numbers.
If we consider each game as a dot and each player as a side of the triangle, then the total number of games (45) would correspond to the 9th triangular number because 1+2+3+4+5+6+7+8+9=45. However, since each side of the triangle represents a player, and there's one more player than sides in our model (because the last player has no one left to play against), we need to add one more to get the total number of players.
Final answer: There were 10 players in the tournament.
Solution 3:
We could also use trial and error to solve this problem. Starting from a small number of players, we calculate how many games would be played in a round-robin tournament until we reach 45 games. For example, with 2 players, only 1 game is played; with 3 players, 3 games are played; with 4 players, 6 games are played, and so on. Continuing this process, we find that with 10 players, 45 games are played.
Final answer: There were 10 players in the tournament.
Ranking of solutions from best to worst:
1. Solution 1: This solution uses direct mathematical reasoning based on the properties of a round-robin tournament and solves an equation to find the answer.
2. Solution 2: This solution uses the concept of triangular numbers which requires additional steps and knowledge beyond the immediate problem context, but still arrives at the correct answer.
3. Solution 3: This solution relies on trial and error, which is less efficient and reliable than the other methods, but it eventually gets to the right answer.
Best and final answer: There were 10 players in the tournament.
```
#### reWOO style execution planning
The model now supports execution planning for complex instructions that would require making use of several tools. The output is just the plan, you must implement a mechanism to parse the output and actually call the functions!
Example prompt:
```
Please construct a systematic plan to generate an optimal response to the user instruction, utilizing a set of provided tools. Each plan will correspond to an evidence value, which will be the output of one of the available functions given an input string
that could be the user's question, one or more prior evidence values, or a combination of both.
Here are the tools available to be called:
DuckDuckGo[input]: Tool that enables the user to search for information using the DuckDuckGo search engine. This tool is helpful in locating a suitable list of sites that may contain the answer to the user's query. It does not directly answer the question but finds a list of sites that might have the answer.
TextScraper[input]: Load one or more websites from the input string containing newline delimited links, where input is one or more links, and produces plain text output containing the content of the links.
HyperlinkExtractor[input]: Extract links from plain text and produces a plain text, newline delimited response of links.
KnowledgeModel[input]: Question answering language model, particularly useful in answering questions based on an input passage of text. The input must be a text question that references an :evidence[n]: variable, e.g. What color is the cat, given :evidence1:?
The output should be in format:
Plan: [first action to take based on input question]
:evidence0: = [function to call with input parameter]
Plan: [next action to take, based on result of :evidence0:]
:evidence1: = [next function to call and input parameter, which may include reference to previous evidence, e.g. "Given context :evidence0"]
...
Answer: [:evidence[n]: containing the final answer.]
Question: Who won the Nobel Prize in Literature in 2020 and what is their most famous work?
```
Response:
```
Plan: Begin by conducting a web search to find out who won the Nobel Prize in Literature in 2020.
:evidence0: = DuckDuckGo[Nobel Prize in Literature 2020 winner]
Plan: Utilize the KnowledgeModel to interpret the search results and find the name of the laureate.
:evidence1: = KnowledgeModel[Who won the Nobel Prize in Literature in 2020, given :evidence0:?]
Plan: Conduct another web search to find the most famous work of the identified laureate.
:evidence2: = DuckDuckGo[Most famous work of :evidence1:]
Plan: Extract the relevant links from the DuckDuckGo search results for a more focused search.
:evidence3: = HyperlinkExtractor[:evidence2:]
Plan: Use the TextScraper tool to extract information from the relevant links.
:evidence4: = TextScraper[:evidence3:]
Plan: Finally, utilize the KnowledgeModel to identify and summarize the most famous work of the laureate from the extracted information.
:evidence5: = KnowledgeModel[What is the most famous work of :evidence1:, given :evidence4:?]
Answer: :evidence5:
```
For this to be useful, you'd have to parse the output plan text, and implement/call each of the functions. This is just pseudo-code, completely untested off the top of my head, and obviously would requiring full implementation + hardening:
```python
import re
import requests
def inject_context(input_text, **context):
for ref in set(re.findall(r"(:evidence[0-9]+:)", input_text, re.I)):
input_text = input_text.replace(ref, context.get(ref, ""))
return input_text
def duckduckgo(input_text, **context):
search_string = inject_context(input_text, **context)
... search via duck duck go using search_string
... return text content
def link_extractor(input_text, **context):
input_text = inject_context(input_text, **context)
return "\n".join(list(set(re.findall(r"(https?://[^\s]+?\.?)", input_text, re.I))))
def scrape(input_text, **context):
input_text = inject_context(input_text, **context)
text = []
for link in input_text.splitlines():
text.append(requests.get(link).text)
return "\n".join(text)
def infer(input_text, **context)
prompt = inject_context(input_text, **context)
... call model with prompt, return output
def parse_plan(plan):
method_map = {
"DuckDuckGo": duckduckgo,
"HyperlinkExtractor": link_extractor,
"KnowledgeModel": infer,
"TextScraper": scrape,
}
context = {}
for line in plan.strip().splitlines():
if line.startswith("Plan:"):
print(line)
continue
parts = re.match("^(:evidence[0-9]+:)\s*=\s*([^\[]+])(\[.*\])\s$", line, re.I)
if not parts:
if line.startswith("Answer: "):
return context.get(line.split(" ")[-1].strip(), "Answer couldn't be generated...")
raise RuntimeError("bad format: " + line)
context[parts.group(1)] = method_map[parts.group(2)](parts.group(3), **context)
```
### Contribute
If you're interested in new functionality, particularly a new "instructor" type to generate a specific type of training data,
take a look at the dataset generation tool repo: https://github.com/jondurbin/airoboros and either make a PR or open an issue with details.
To help me with the OpenAI/compute costs:
- https://bmc.link/jondurbin
- ETH 0xce914eAFC2fe52FdceE59565Dd92c06f776fcb11
- BTC bc1qdwuth4vlg8x37ggntlxu5cjfwgmdy5zaa7pswf
### Licence and usage restrictions
The airoboros 3.0 models are built on top of multiple base models, each with their own license/restrictions.
The models with `-l2` in the name have a custom Meta license:
- See the [meta-license/LICENSE.txt](meta-license/LICENSE.txt) file attached for the original license provided by Meta.
- See also [meta-license/USE_POLICY.md](meta-license/USE_POLICY.md) and [meta-license/Responsible-Use-Guide.pdf](meta-license/Responsible-Use-Guide.pdf), also provided by Meta.
The models with `-m-` are mistral-7b (apache 2.0)
The model with `-3b` uses Stability AI, which as a `cc-by-sa-4.0` license.
The fine-tuning data was mostly generated by OpenAI API calls to gpt-4, via [airoboros](https://github.com/jondurbin/airoboros)
The ToS for OpenAI API usage has a clause preventing the output from being used to train a model that __competes__ with OpenAI
- what does *compete* actually mean here?
- these small open source models will not produce output anywhere near the quality of gpt-4, or even gpt-3.5, so I can't imagine this could credibly be considered competing in the first place
- if someone else uses the dataset to do the same, they wouldn't necessarily be violating the ToS because they didn't call the API, so I don't know how that works
- the training data used in essentially all large language models includes a significant amount of copyrighted or otherwise non-permissive licensing in the first place
- other work using the self-instruct method, e.g. the original here: https://github.com/yizhongw/self-instruct released the data and model as apache-2
I am purposingly leaving this license ambiguous (other than the fact you must comply with the Meta original license for llama-2) because I am not a lawyer and refuse to attempt to interpret all of the terms accordingly.
Your best bet is probably to avoid using this commercially due to the OpenAI API usage.
Either way, by using this model, you agree to completely indemnify me.
<!-- original-model-card end -->
|
TheBloke/airoboros-l2-13B-3.0-GPTQ
|
TheBloke
| 2023-10-05T23:09:23Z | 25 | 1 |
transformers
|
[
"transformers",
"safetensors",
"llama",
"text-generation",
"dataset:jondurbin/airoboros-3.0",
"base_model:jondurbin/airoboros-l2-13b-3.0",
"base_model:quantized:jondurbin/airoboros-l2-13b-3.0",
"license:llama2",
"autotrain_compatible",
"text-generation-inference",
"4-bit",
"gptq",
"region:us"
] |
text-generation
| 2023-10-05T22:38:15Z |
---
base_model: jondurbin/airoboros-l2-13b-3.0
datasets:
- jondurbin/airoboros-3.0
inference: false
license: llama2
model_creator: Jon Durbin
model_name: Airoboros L2 13B 3.0
model_type: llama
prompt_template: '[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
'
quantized_by: TheBloke
---
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Airoboros L2 13B 3.0 - GPTQ
- Model creator: [Jon Durbin](https://huggingface.co/jondurbin)
- Original model: [Airoboros L2 13B 3.0](https://huggingface.co/jondurbin/airoboros-l2-13b-3.0)
<!-- description start -->
## Description
This repo contains GPTQ model files for [Jon Durbin's Airoboros L2 13B 3.0](https://huggingface.co/jondurbin/airoboros-l2-13b-3.0).
Multiple GPTQ parameter permutations are provided; see Provided Files below for details of the options provided, their parameters, and the software used to create them.
<!-- description end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GGUF)
* [Jon Durbin's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/jondurbin/airoboros-l2-13b-3.0)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: Llama-2-Uncensored
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
```
<!-- prompt-template end -->
<!-- README_GPTQ.md-provided-files start -->
## Provided files, and GPTQ parameters
Multiple quantisation parameters are provided, to allow you to choose the best one for your hardware and requirements.
Each separate quant is in a different branch. See below for instructions on fetching from different branches.
Most GPTQ files are made with AutoGPTQ. Mistral models are currently made with Transformers.
<details>
<summary>Explanation of GPTQ parameters</summary>
- Bits: The bit size of the quantised model.
- GS: GPTQ group size. Higher numbers use less VRAM, but have lower quantisation accuracy. "None" is the lowest possible value.
- Act Order: True or False. Also known as `desc_act`. True results in better quantisation accuracy. Some GPTQ clients have had issues with models that use Act Order plus Group Size, but this is generally resolved now.
- Damp %: A GPTQ parameter that affects how samples are processed for quantisation. 0.01 is default, but 0.1 results in slightly better accuracy.
- GPTQ dataset: The calibration dataset used during quantisation. Using a dataset more appropriate to the model's training can improve quantisation accuracy. Note that the GPTQ calibration dataset is not the same as the dataset used to train the model - please refer to the original model repo for details of the training dataset(s).
- Sequence Length: The length of the dataset sequences used for quantisation. Ideally this is the same as the model sequence length. For some very long sequence models (16+K), a lower sequence length may have to be used. Note that a lower sequence length does not limit the sequence length of the quantised model. It only impacts the quantisation accuracy on longer inference sequences.
- ExLlama Compatibility: Whether this file can be loaded with ExLlama, which currently only supports Llama models in 4-bit.
</details>
| Branch | Bits | GS | Act Order | Damp % | GPTQ Dataset | Seq Len | Size | ExLlama | Desc |
| ------ | ---- | -- | --------- | ------ | ------------ | ------- | ---- | ------- | ---- |
| [main](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GPTQ/tree/main) | 4 | 128 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 7.26 GB | Yes | 4-bit, with Act Order and group size 128g. Uses even less VRAM than 64g, but with slightly lower accuracy. |
| [gptq-4bit-32g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GPTQ/tree/gptq-4bit-32g-actorder_True) | 4 | 32 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 8.00 GB | Yes | 4-bit, with Act Order and group size 32g. Gives highest possible inference quality, with maximum VRAM usage. |
| [gptq-8bit--1g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GPTQ/tree/gptq-8bit--1g-actorder_True) | 8 | None | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 13.36 GB | No | 8-bit, with Act Order. No group size, to lower VRAM requirements. |
| [gptq-8bit-128g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GPTQ/tree/gptq-8bit-128g-actorder_True) | 8 | 128 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 13.65 GB | No | 8-bit, with group size 128g for higher inference quality and with Act Order for even higher accuracy. |
| [gptq-8bit-32g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GPTQ/tree/gptq-8bit-32g-actorder_True) | 8 | 32 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 14.54 GB | No | 8-bit, with group size 32g and Act Order for maximum inference quality. |
| [gptq-4bit-64g-actorder_True](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GPTQ/tree/gptq-4bit-64g-actorder_True) | 4 | 64 | Yes | 0.1 | [wikitext](https://huggingface.co/datasets/wikitext/viewer/wikitext-2-v1/test) | 4096 | 7.51 GB | Yes | 4-bit, with Act Order and group size 64g. Uses less VRAM than 32g, but with slightly lower accuracy. |
<!-- README_GPTQ.md-provided-files end -->
<!-- README_GPTQ.md-download-from-branches start -->
## How to download, including from branches
### In text-generation-webui
To download from the `main` branch, enter `TheBloke/airoboros-l2-13B-3.0-GPTQ` in the "Download model" box.
To download from another branch, add `:branchname` to the end of the download name, eg `TheBloke/airoboros-l2-13B-3.0-GPTQ:gptq-4bit-32g-actorder_True`
### From the command line
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
To download the `main` branch to a folder called `airoboros-l2-13B-3.0-GPTQ`:
```shell
mkdir airoboros-l2-13B-3.0-GPTQ
huggingface-cli download TheBloke/airoboros-l2-13B-3.0-GPTQ --local-dir airoboros-l2-13B-3.0-GPTQ --local-dir-use-symlinks False
```
To download from a different branch, add the `--revision` parameter:
```shell
mkdir airoboros-l2-13B-3.0-GPTQ
huggingface-cli download TheBloke/airoboros-l2-13B-3.0-GPTQ --revision gptq-4bit-32g-actorder_True --local-dir airoboros-l2-13B-3.0-GPTQ --local-dir-use-symlinks False
```
<details>
<summary>More advanced huggingface-cli download usage</summary>
If you remove the `--local-dir-use-symlinks False` parameter, the files will instead be stored in the central Huggingface cache directory (default location on Linux is: `~/.cache/huggingface`), and symlinks will be added to the specified `--local-dir`, pointing to their real location in the cache. This allows for interrupted downloads to be resumed, and allows you to quickly clone the repo to multiple places on disk without triggering a download again. The downside, and the reason why I don't list that as the default option, is that the files are then hidden away in a cache folder and it's harder to know where your disk space is being used, and to clear it up if/when you want to remove a download model.
The cache location can be changed with the `HF_HOME` environment variable, and/or the `--cache-dir` parameter to `huggingface-cli`.
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
mkdir airoboros-l2-13B-3.0-GPTQ
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/airoboros-l2-13B-3.0-GPTQ --local-dir airoboros-l2-13B-3.0-GPTQ --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
</details>
### With `git` (**not** recommended)
To clone a specific branch with `git`, use a command like this:
```shell
git clone --single-branch --branch gptq-4bit-32g-actorder_True https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GPTQ
```
Note that using Git with HF repos is strongly discouraged. It will be much slower than using `huggingface-hub`, and will use twice as much disk space as it has to store the model files twice (it stores every byte both in the intended target folder, and again in the `.git` folder as a blob.)
<!-- README_GPTQ.md-download-from-branches end -->
<!-- README_GPTQ.md-text-generation-webui start -->
## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
It is strongly recommended to use the text-generation-webui one-click-installers unless you're sure you know how to make a manual install.
1. Click the **Model tab**.
2. Under **Download custom model or LoRA**, enter `TheBloke/airoboros-l2-13B-3.0-GPTQ`.
- To download from a specific branch, enter for example `TheBloke/airoboros-l2-13B-3.0-GPTQ:gptq-4bit-32g-actorder_True`
- see Provided Files above for the list of branches for each option.
3. Click **Download**.
4. The model will start downloading. Once it's finished it will say "Done".
5. In the top left, click the refresh icon next to **Model**.
6. In the **Model** dropdown, choose the model you just downloaded: `airoboros-l2-13B-3.0-GPTQ`
7. The model will automatically load, and is now ready for use!
8. If you want any custom settings, set them and then click **Save settings for this model** followed by **Reload the Model** in the top right.
* Note that you do not need to and should not set manual GPTQ parameters any more. These are set automatically from the file `quantize_config.json`.
9. Once you're ready, click the **Text Generation tab** and enter a prompt to get started!
<!-- README_GPTQ.md-text-generation-webui end -->
<!-- README_GPTQ.md-use-from-tgi start -->
## Serving this model from Text Generation Inference (TGI)
It's recommended to use TGI version 1.1.0 or later. The official Docker container is: `ghcr.io/huggingface/text-generation-inference:1.1.0`
Example Docker parameters:
```shell
--model-id TheBloke/airoboros-l2-13B-3.0-GPTQ --port 3000 --quantize awq --max-input-length 3696 --max-total-tokens 4096 --max-batch-prefill-tokens 4096
```
Example Python code for interfacing with TGI (requires huggingface-hub 0.17.0 or later):
```shell
pip3 install huggingface-hub
```
```python
from huggingface_hub import InferenceClient
endpoint_url = "https://your-endpoint-url-here"
prompt = "Tell me about AI"
prompt_template=f'''[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
'''
client = InferenceClient(endpoint_url)
response = client.text_generation(prompt,
max_new_tokens=128,
do_sample=True,
temperature=0.7,
top_p=0.95,
top_k=40,
repetition_penalty=1.1)
print(f"Model output: {response}")
```
<!-- README_GPTQ.md-use-from-tgi end -->
<!-- README_GPTQ.md-use-from-python start -->
## How to use this GPTQ model from Python code
### Install the necessary packages
Requires: Transformers 4.33.0 or later, Optimum 1.12.0 or later, and AutoGPTQ 0.4.2 or later.
```shell
pip3 install transformers optimum
pip3 install auto-gptq --extra-index-url https://huggingface.github.io/autogptq-index/whl/cu118/ # Use cu117 if on CUDA 11.7
```
If you have problems installing AutoGPTQ using the pre-built wheels, install it from source instead:
```shell
pip3 uninstall -y auto-gptq
git clone https://github.com/PanQiWei/AutoGPTQ
cd AutoGPTQ
git checkout v0.4.2
pip3 install .
```
### You can then use the following code
```python
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
model_name_or_path = "TheBloke/airoboros-l2-13B-3.0-GPTQ"
# To use a different branch, change revision
# For example: revision="gptq-4bit-32g-actorder_True"
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
device_map="auto",
trust_remote_code=False,
revision="main")
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
prompt = "Tell me about AI"
prompt_template=f'''[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, do_sample=True, top_p=0.95, top_k=40, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
do_sample=True,
temperature=0.7,
top_p=0.95,
top_k=40,
repetition_penalty=1.1
)
print(pipe(prompt_template)[0]['generated_text'])
```
<!-- README_GPTQ.md-use-from-python end -->
<!-- README_GPTQ.md-compatibility start -->
## Compatibility
The files provided are tested to work with AutoGPTQ, both via Transformers and using AutoGPTQ directly. They should also work with [Occ4m's GPTQ-for-LLaMa fork](https://github.com/0cc4m/KoboldAI).
[ExLlama](https://github.com/turboderp/exllama) is compatible with Llama and Mistral models in 4-bit. Please see the Provided Files table above for per-file compatibility.
[Huggingface Text Generation Inference (TGI)](https://github.com/huggingface/text-generation-inference) is compatible with all GPTQ models.
<!-- README_GPTQ.md-compatibility end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Pierre Kircher, Stanislav Ovsiannikov, Michael Levine, Eugene Pentland, Andrey, 준교 김, Randy H, Fred von Graf, Artur Olbinski, Caitlyn Gatomon, terasurfer, Jeff Scroggin, James Bentley, Vadim, Gabriel Puliatti, Harry Royden McLaughlin, Sean Connelly, Dan Guido, Edmond Seymore, Alicia Loh, subjectnull, AzureBlack, Manuel Alberto Morcote, Thomas Belote, Lone Striker, Chris Smitley, Vitor Caleffi, Johann-Peter Hartmann, Clay Pascal, biorpg, Brandon Frisco, sidney chen, transmissions 11, Pedro Madruga, jinyuan sun, Ajan Kanaga, Emad Mostaque, Trenton Dambrowitz, Jonathan Leane, Iucharbius, usrbinkat, vamX, George Stoitzev, Luke Pendergrass, theTransient, Olakabola, Swaroop Kallakuri, Cap'n Zoog, Brandon Phillips, Michael Dempsey, Nikolai Manek, danny, Matthew Berman, Gabriel Tamborski, alfie_i, Raymond Fosdick, Tom X Nguyen, Raven Klaugh, LangChain4j, Magnesian, Illia Dulskyi, David Ziegler, Mano Prime, Luis Javier Navarrete Lozano, Erik Bjäreholt, 阿明, Nathan Dryer, Alex, Rainer Wilmers, zynix, TL, Joseph William Delisle, John Villwock, Nathan LeClaire, Willem Michiel, Joguhyik, GodLy, OG, Alps Aficionado, Jeffrey Morgan, ReadyPlayerEmma, Tiffany J. Kim, Sebastain Graf, Spencer Kim, Michael Davis, webtim, Talal Aujan, knownsqashed, John Detwiler, Imad Khwaja, Deo Leter, Jerry Meng, Elijah Stavena, Rooh Singh, Pieter, SuperWojo, Alexandros Triantafyllidis, Stephen Murray, Ai Maven, ya boyyy, Enrico Ros, Ken Nordquist, Deep Realms, Nicholas, Spiking Neurons AB, Elle, Will Dee, Jack West, RoA, Luke @flexchar, Viktor Bowallius, Derek Yates, Subspace Studios, jjj, Toran Billups, Asp the Wyvern, Fen Risland, Ilya, NimbleBox.ai, Chadd, Nitin Borwankar, Emre, Mandus, Leonard Tan, Kalila, K, Trailburnt, S_X, Cory Kujawski
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
# Original model card: Jon Durbin's Airoboros L2 13B 3.0
### Overview
Another experimental model, using mostly sythetic data generated by [airoboros](https://github.com/jondurbin/airoboros)
#### IMPORTANT NOTE - llama-2 chat format!!!
The airobors 3.0 models have been updated to use llama-2 chat format, rather than vicuna style user/assistant!
This is a breaking change, although most inference systems support llama-2 chat templating.
See prompt format below for all of the details.
#### Highlights
A model built on [airoboros-3.0 dataset](https://hf.co/datasets/jondurbin/airoboros-3.0), which is a continuation of [airoboros-2.2.1 dataset](https://hf.co/datasets/jondurbin/airoboros-2.2.1) with the following extras:
* MathJSON - this is not yet part of the code directly, but can be easily parsed and executed, as a way to perform deterministic calculations. LLMs suck at math, so this does problem formulation and let's the work be done by traditional operations.
* Human-generated RP multi-turn conversation - provided by an anon, this is a human created RP dataset, meant to assist with multi-turn coherency and chat experience.
This is a fairly general purpose model, but focuses heavily on instruction following, rather than casual chat/roleplay.
Huge thank you to the folks over at [a16z](https://a16z.com/) for sponsoring the costs associated with building models and associated tools!
### Prompt format
The prompt format is llama-2 chat.
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
```
For multi-turn, the prompt format is as follows:
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt 0} [/INST] {response 0} </s><s>[INST] {prompt 1} [/INST] {response 1} </s><s>...[INST] {prompt N} [/INST]
```
Here is a simple python script that can help generate the prompt:
```python
def get_chat_prompt(message, chat_history, system_prompt):
texts = [f'[INST] <<SYS>>\n{system_prompt}\n<</SYS>>\n\n']
do_strip = False
for user_input, response in chat_history:
user_input = user_input.strip() if do_strip else user_input
do_strip = True
texts.append(f'{user_input} [/INST] {response.strip()} </s><s>[INST] ')
message = message.strip() if do_strip else message
texts.append(f'{message} [/INST] ')
return ''.join(texts)
def get_prompt(item):
"""
`item` is expected to be in chatml format, e.g.
{
"id": 1,
"conversations": [
{
"role": "system",
"content": "You are a helpful, unbiased, uncensored assistant."
},
{
"role": "user",
"content": "Hi!"
},
{
"role": "assistant",
"content": "Hello, how are you today?"
},
{
"role": "user",
"content": "Just peachy, and you?"
}
]
}
"""
system_prompt = share_gpt_item['conversations'][0]['value']
offset = 1
if share_gpt_item['conversations'][0]['role'] != 'system':
system_prompt = 'You are an unbiased, uncensored, helpful assistant.'
offset = 0
chat_history = [
(share_gpt_item['conversations'][idx]['content'], share_gpt_item['conversations'][idx + 1]['content'])
for idx in range(offset, len(item['conversations']) - 1, 2)
]
message = item['conversations'][-1]['content']
return get_chat_prompt(message, chat_history, system_prompt)
```
### Fine tuning info
https://wandb.ai/jondurbin/airoboros-l2-13b-3.0/runs/97utxyk9?workspace=user-jondurbin
### Helpful usage tips
#### MathJSON
Ask mathematical questions, prefixed with "Create a MathJSON solution to the following:", and you should get output with a formula that can be executed by https://cortexjs.io/compute-engine/
I also made a (really quite hacky, incomplete, semi-functional) python version that works for most expressions, see [mathjson.py](https://github.com/jondurbin/airoboros/blob/mathjson/airoboros/mathjson.py)
__Use a very low temperature!__
Example:
```
[INST] <<SYS>>
You are an assistant with vast knowledge in all things math.
<</SYS>
Create a MathJSON solution to the following: Calculate the area of a circle with radius of 17.2456cm. Include your reasoning. [/INST]
```
Output:
```
The area (A) of a circle is given by the formula A = πr², where r is the radius of the circle. Since the radius is given in centimeters, we need to convert it to meters first. We know that 1 meter equals 100 centimeters. Therefore, we can represent the radius as: r = 17.2456 cm * 100/1.
Once we have the radius in meters, we can calculate the area of the circle using the formula A = πr².
Solution as MathJSON:
<mathjson>
[
"Multiply",
"Pi",
[
"Power",
17.2456,
2
]
]
</mathjson>
```
You can then validate the JSON between `<mathjson>` and `</mathjson>`, then pass the parsed JSON to compute-engine JS or the `evaluate` function in mathjson.py to calculate the response.
#### Context obedient question answering
By obedient, I mean the model was trained to ignore what it thinks it knows, and uses the context to answer the question. The model was also tuned to limit the values to the provided context as much as possible to reduce hallucinations.
The format for a closed-context prompt is as follows:
```
BEGININPUT
BEGINCONTEXT
[key0: value0]
[key1: value1]
... other metdata ...
ENDCONTEXT
[insert your text blocks here]
ENDINPUT
[add as many other blocks, in the exact same format]
BEGININSTRUCTION
[insert your instruction(s). The model was tuned with single questions, paragraph format, lists, etc.]
ENDINSTRUCTION
```
It's also helpful to add "Don't make up answers if you don't know." to your instruction block to make sure if the context is completely unrelated it doesn't make something up.
*The __only__ prompts that need this closed context formating are closed-context instructions. Normal questions/instructions do not!*
I know it's a bit verbose and annoying, but after much trial and error, using these explicit delimiters helps the model understand where to find the responses and how to associate specific sources with it.
- `BEGININPUT` - denotes a new input block
- `BEGINCONTEXT` - denotes the block of context (metadata key/value pairs) to associate with the current input block
- `ENDCONTEXT` - denotes the end of the metadata block for the current input
- [text] - Insert whatever text you want for the input block, as many paragraphs as can fit in the context.
- `ENDINPUT` - denotes the end of the current input block
- [repeat as many input blocks in this format as you want]
- `BEGININSTRUCTION` - denotes the start of the list (or one) instruction(s) to respond to for all of the input blocks above.
- [instruction(s)]
- `ENDINSTRUCTION` - denotes the end of instruction set
It sometimes works without `ENDINSTRUCTION`, but by explicitly including that in the prompt, the model better understands that all of the instructions in the block should be responded to.
__Use a very low temperature!__
Here's a trivial, but important example to prove the point:
```
BEGININPUT
BEGINCONTEXT
date: 2021-01-01
url: https://web.site/123
ENDCONTEXT
In a shocking turn of events, blueberries are now green, but will be sticking with the same name.
ENDINPUT
BEGININSTRUCTION
What color are bluberries? Source?
ENDINSTRUCTION
```
And the response:
```
Blueberries are now green.
Source:
date: 2021-01-01
url: https://web.site/123
```
#### Summarization
500 samples have been included from [this dataset](https://huggingface.co/datasets/mattpscott/airoboros-summarization), using the same format as contextual question answering, for example:
```
BEGININPUT
{text to summarize}
ENDINPUT
BEGININSTRUCTION
Summarize the input in around 130 words.
ENDINSTRUCTION
```
#### Getting longer responses
You can use a few techniques to get longer responses.
Detailed prompts, with explicit instruction for word count:
```
Please compose a narrative set in the heart of an ancient library, steeped in the scent of old parchment and ink. The protagonist should be a young scholar who is dedicated to studying the art of storytelling and its evolution throughout history. In her pursuit of knowledge, she stumbles upon a forgotten tome that seems to possess an unusual aura. This book has the ability to bring stories to life, literally manifesting characters and scenarios from within its pages into reality.
The main character must navigate through various epochs of storytelling - from oral traditions of tribal societies, through medieval minstrels' tales, to modern-day digital narratives - as they come alive around her. Each era presents its unique challenges and lessons about the power and impact of stories on human civilization.
One such character could be a sentient quill pen, who was once used by renowned authors of yesteryears and now holds their wisdom and experiences. It becomes her mentor, guiding her through this journey with witty remarks and insightful commentary.
Ensure that your tale encapsulates the thrill of adventure, the beauty of learning, and the profound connection between humans and their stories. All characters involved should be non-human entities. Feel free to explore creative liberties but maintain the mentioned elements.
Your response should be approximately 2300 words.
```
Or, a simpler example:
```
Please create a long, detailed story about a dragon in an old growth forest who, for some reason, begins speaking the words of the source code of linux.
```
There are a few examples of next chapter completion as well, e.g.:
```
Write the next chapter of a historical fiction novel set in Paris during the 20th century.
Here's a summary of the previous chapter:
In the vibrant city of Paris, amid the tumultuous changes of the 20th century, our protagonist Margot, an aspiring fashion designer, has just secured an apprenticeship at a prestigious couture house. She meets Lucien, a charming journalist who covers the fashion industry. Together they navigate the ever-changing world of fashion and society, uncovering secrets that reveal the intricate links between style, politics, and culture. As the chapter concludes, they decide to delve deeper into the hidden corners of the fashion world to unravel its mysteries.
Requirements for the next chapter:
1. Character Development of Margot and Lucien:
- Margot's Evolution: Unfold more about Margot's past, her dreams of revolutionizing fashion, and her struggle to establish herself in a male-dominated industry. Illustrate her growing expertise, innovative ideas, and increasing dependence on Lucien.
- Lucien's Complexity: Introduce uncertainties surrounding Lucien's background and real motives. Increase suspense by suggesting undisclosed information he possesses, while also highlighting his wit and perceptiveness.
2. Exploration of Paris and the Couture House:
- Paris: Elaborate their journey through the bustling streets of Paris, including encounters with iconic figures, social unrest, and relics from different eras of French history.
- The Couture House: Expand on the grandeur of the couture house they work in, filled with artistic masterpieces, intense competition, and cryptic notes hinting at a scandalous past.
3. Emergence of the Subplot: The Lost Collection:
- Discovery: Have Margot and Lucien stumble upon a secret vault containing a lost collection designed before World War II, raising new questions about the previous owner and the influence of war on fashion.
- Revelation: Capture their shock as they realize the designs were plagiarized, the potential repercussions, and the opportunities it presents for Margot's career.
- Twist: End with a twist that suggests there are other stolen collections across Paris, setting up their new mission.
Your response should be approximately 650 words.
```
#### Coding
You can ask for fairly complex coding instructions with multiple criteria, e.g.:
```
Create a python application with the following requirements:
- Asyncio FastAPI webserver
- ping endpoint that returns the current date in JSON format
- file upload endpoint, which calculates the file's sha256 checksum, and checks postgres to deduplicate
```
Or inline criteria:
```
Write a multi-threaded TCP server in C that accepts a "GET [key]" input and "SET [key] [value]" input, and uses a binary tree to get and store the input values.
```
You can also optionally add a single space and "PLAINFORMAT" at the end of your prompt to avoid backticks, explanations, etc. and just print the code, e.g.:
```
Write a websocket application in node.js. PLAINFORMAT
```
#### Agent/function calling
The dataset includes many examples of function/args generation based on input criteria. This is somewhat similar to the OpenAI function calling, but the output is either JSON or YAML.
Example prompt:
```
As an AI assistant, please select the most suitable function and parameters from the list of available functions below, based on the user's input. Provide your response in JSON format.
Input: I want to know how many times 'Python' is mentioned in my text file.
Available functions:
file_analytics:
description: This tool performs various operations on a text file.
params:
action: The operation we want to perform on the data, such as "count_occurrences", "find_line", etc.
filters:
keyword: The word or phrase we want to search for.
```
Response:
```json
{
"function": "file_analytics",
"params": {
"action": "count_occurrences",
"filters": {
"keyword": "Python"
}
}
}
```
#### Chain-of-thought
You can ask for several possible responses to a given problem, with a ranking and final answer selection.
Example prompt:
```
A group of people decide to play a round-robin tournament where every player plays every other player exactly once. If a total of 45 games were played, how many players participated in the tournament? Offer several potential answers, rank them according to mathematical logic, and pick the most feasible one.
```
Example response:
```
Solution 1:
In a round-robin tournament, each player plays every other player exactly once. This means that the total number of games played is equal to the combination of all players taken two at a time (since a game involves two players).
The formula for combinations is nC2 = n(n-1)/2, where n is the total number of players. We know that nC2 equals 45 (the total number of games), so we can set up the equation as follows:
n(n-1)/2 = 45
Solving this equation gives us n=10.
Final answer: There were 10 players in the tournament.
Solution 2:
Another way to approach this problem is by using the concept of triangular numbers. Triangular numbers are a sequence of numbers where each number can be represented as a triangle with dots. The nth triangular number is given by the sum of the first n natural numbers.
If we consider each game as a dot and each player as a side of the triangle, then the total number of games (45) would correspond to the 9th triangular number because 1+2+3+4+5+6+7+8+9=45. However, since each side of the triangle represents a player, and there's one more player than sides in our model (because the last player has no one left to play against), we need to add one more to get the total number of players.
Final answer: There were 10 players in the tournament.
Solution 3:
We could also use trial and error to solve this problem. Starting from a small number of players, we calculate how many games would be played in a round-robin tournament until we reach 45 games. For example, with 2 players, only 1 game is played; with 3 players, 3 games are played; with 4 players, 6 games are played, and so on. Continuing this process, we find that with 10 players, 45 games are played.
Final answer: There were 10 players in the tournament.
Ranking of solutions from best to worst:
1. Solution 1: This solution uses direct mathematical reasoning based on the properties of a round-robin tournament and solves an equation to find the answer.
2. Solution 2: This solution uses the concept of triangular numbers which requires additional steps and knowledge beyond the immediate problem context, but still arrives at the correct answer.
3. Solution 3: This solution relies on trial and error, which is less efficient and reliable than the other methods, but it eventually gets to the right answer.
Best and final answer: There were 10 players in the tournament.
```
#### reWOO style execution planning
The model now supports execution planning for complex instructions that would require making use of several tools. The output is just the plan, you must implement a mechanism to parse the output and actually call the functions!
Example prompt:
```
Please construct a systematic plan to generate an optimal response to the user instruction, utilizing a set of provided tools. Each plan will correspond to an evidence value, which will be the output of one of the available functions given an input string
that could be the user's question, one or more prior evidence values, or a combination of both.
Here are the tools available to be called:
DuckDuckGo[input]: Tool that enables the user to search for information using the DuckDuckGo search engine. This tool is helpful in locating a suitable list of sites that may contain the answer to the user's query. It does not directly answer the question but finds a list of sites that might have the answer.
TextScraper[input]: Load one or more websites from the input string containing newline delimited links, where input is one or more links, and produces plain text output containing the content of the links.
HyperlinkExtractor[input]: Extract links from plain text and produces a plain text, newline delimited response of links.
KnowledgeModel[input]: Question answering language model, particularly useful in answering questions based on an input passage of text. The input must be a text question that references an :evidence[n]: variable, e.g. What color is the cat, given :evidence1:?
The output should be in format:
Plan: [first action to take based on input question]
:evidence0: = [function to call with input parameter]
Plan: [next action to take, based on result of :evidence0:]
:evidence1: = [next function to call and input parameter, which may include reference to previous evidence, e.g. "Given context :evidence0"]
...
Answer: [:evidence[n]: containing the final answer.]
Question: Who won the Nobel Prize in Literature in 2020 and what is their most famous work?
```
Response:
```
Plan: Begin by conducting a web search to find out who won the Nobel Prize in Literature in 2020.
:evidence0: = DuckDuckGo[Nobel Prize in Literature 2020 winner]
Plan: Utilize the KnowledgeModel to interpret the search results and find the name of the laureate.
:evidence1: = KnowledgeModel[Who won the Nobel Prize in Literature in 2020, given :evidence0:?]
Plan: Conduct another web search to find the most famous work of the identified laureate.
:evidence2: = DuckDuckGo[Most famous work of :evidence1:]
Plan: Extract the relevant links from the DuckDuckGo search results for a more focused search.
:evidence3: = HyperlinkExtractor[:evidence2:]
Plan: Use the TextScraper tool to extract information from the relevant links.
:evidence4: = TextScraper[:evidence3:]
Plan: Finally, utilize the KnowledgeModel to identify and summarize the most famous work of the laureate from the extracted information.
:evidence5: = KnowledgeModel[What is the most famous work of :evidence1:, given :evidence4:?]
Answer: :evidence5:
```
For this to be useful, you'd have to parse the output plan text, and implement/call each of the functions. This is just pseudo-code, completely untested off the top of my head, and obviously would requiring full implementation + hardening:
```python
import re
import requests
def inject_context(input_text, **context):
for ref in set(re.findall(r"(:evidence[0-9]+:)", input_text, re.I)):
input_text = input_text.replace(ref, context.get(ref, ""))
return input_text
def duckduckgo(input_text, **context):
search_string = inject_context(input_text, **context)
... search via duck duck go using search_string
... return text content
def link_extractor(input_text, **context):
input_text = inject_context(input_text, **context)
return "\n".join(list(set(re.findall(r"(https?://[^\s]+?\.?)", input_text, re.I))))
def scrape(input_text, **context):
input_text = inject_context(input_text, **context)
text = []
for link in input_text.splitlines():
text.append(requests.get(link).text)
return "\n".join(text)
def infer(input_text, **context)
prompt = inject_context(input_text, **context)
... call model with prompt, return output
def parse_plan(plan):
method_map = {
"DuckDuckGo": duckduckgo,
"HyperlinkExtractor": link_extractor,
"KnowledgeModel": infer,
"TextScraper": scrape,
}
context = {}
for line in plan.strip().splitlines():
if line.startswith("Plan:"):
print(line)
continue
parts = re.match("^(:evidence[0-9]+:)\s*=\s*([^\[]+])(\[.*\])\s$", line, re.I)
if not parts:
if line.startswith("Answer: "):
return context.get(line.split(" ")[-1].strip(), "Answer couldn't be generated...")
raise RuntimeError("bad format: " + line)
context[parts.group(1)] = method_map[parts.group(2)](parts.group(3), **context)
```
### Contribute
If you're interested in new functionality, particularly a new "instructor" type to generate a specific type of training data,
take a look at the dataset generation tool repo: https://github.com/jondurbin/airoboros and either make a PR or open an issue with details.
To help me with the OpenAI/compute costs:
- https://bmc.link/jondurbin
- ETH 0xce914eAFC2fe52FdceE59565Dd92c06f776fcb11
- BTC bc1qdwuth4vlg8x37ggntlxu5cjfwgmdy5zaa7pswf
### Licence and usage restrictions
The airoboros 3.0 models are built on top of multiple base models, each with their own license/restrictions.
The models with `-l2` in the name have a custom Meta license:
- See the [meta-license/LICENSE.txt](meta-license/LICENSE.txt) file attached for the original license provided by Meta.
- See also [meta-license/USE_POLICY.md](meta-license/USE_POLICY.md) and [meta-license/Responsible-Use-Guide.pdf](meta-license/Responsible-Use-Guide.pdf), also provided by Meta.
The models with `-m-` are mistral-7b (apache 2.0)
The model with `-3b` uses Stability AI, which as a `cc-by-sa-4.0` license.
The fine-tuning data was mostly generated by OpenAI API calls to gpt-4, via [airoboros](https://github.com/jondurbin/airoboros)
The ToS for OpenAI API usage has a clause preventing the output from being used to train a model that __competes__ with OpenAI
- what does *compete* actually mean here?
- these small open source models will not produce output anywhere near the quality of gpt-4, or even gpt-3.5, so I can't imagine this could credibly be considered competing in the first place
- if someone else uses the dataset to do the same, they wouldn't necessarily be violating the ToS because they didn't call the API, so I don't know how that works
- the training data used in essentially all large language models includes a significant amount of copyrighted or otherwise non-permissive licensing in the first place
- other work using the self-instruct method, e.g. the original here: https://github.com/yizhongw/self-instruct released the data and model as apache-2
I am purposingly leaving this license ambiguous (other than the fact you must comply with the Meta original license for llama-2) because I am not a lawyer and refuse to attempt to interpret all of the terms accordingly.
Your best bet is probably to avoid using this commercially due to the OpenAI API usage.
Either way, by using this model, you agree to completely indemnify me.
|
kaveshan/mistral-7b-kn-instruct
|
kaveshan
| 2023-10-05T23:07:35Z | 0 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-10-05T20:47:15Z |
---
library_name: peft
---
## Training procedure
The following `bitsandbytes` quantization config was used during training:
- quant_method: bitsandbytes
- 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 0.6.0.dev0
|
MaxReynolds/SouderRocketLauncherNet
|
MaxReynolds
| 2023-10-05T22:56:46Z | 4 | 0 |
diffusers
|
[
"diffusers",
"tensorboard",
"stable-diffusion",
"stable-diffusion-diffusers",
"text-to-image",
"lora",
"base_model:CompVis/stable-diffusion-v1-2",
"base_model:adapter:CompVis/stable-diffusion-v1-2",
"license:creativeml-openrail-m",
"region:us"
] |
text-to-image
| 2023-09-30T01:30:43Z |
---
license: creativeml-openrail-m
base_model: CompVis/stable-diffusion-v1-2
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
- diffusers
- lora
inference: true
---
# LoRA text2image fine-tuning - MaxReynolds/SouderRocketLauncherNet
These are LoRA adaption weights for CompVis/stable-diffusion-v1-2. The weights were fine-tuned on the MaxReynolds/Lee_Souder_RocketLauncher dataset. You can find some example images in the following.
|
TheBloke/airoboros-l2-13B-3.0-GGUF
|
TheBloke
| 2023-10-05T22:47:05Z | 110 | 7 |
transformers
|
[
"transformers",
"gguf",
"llama",
"dataset:jondurbin/airoboros-3.0",
"base_model:jondurbin/airoboros-l2-13b-3.0",
"base_model:quantized:jondurbin/airoboros-l2-13b-3.0",
"license:llama2",
"region:us"
] | null | 2023-10-05T22:38:05Z |
---
base_model: jondurbin/airoboros-l2-13b-3.0
datasets:
- jondurbin/airoboros-3.0
inference: false
license: llama2
model_creator: Jon Durbin
model_name: Airoboros L2 13B 3.0
model_type: llama
prompt_template: '[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
'
quantized_by: TheBloke
---
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Airoboros L2 13B 3.0 - GGUF
- Model creator: [Jon Durbin](https://huggingface.co/jondurbin)
- Original model: [Airoboros L2 13B 3.0](https://huggingface.co/jondurbin/airoboros-l2-13b-3.0)
<!-- description start -->
## Description
This repo contains GGUF format model files for [Jon Durbin's Airoboros L2 13B 3.0](https://huggingface.co/jondurbin/airoboros-l2-13b-3.0).
<!-- description end -->
<!-- README_GGUF.md-about-gguf start -->
### About GGUF
GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.
Here is an incomplate list of clients and libraries that are known to support GGUF:
* [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option.
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.
* [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.
* [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration.
* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection.
* [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.
* [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server.
* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.
* [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use.
<!-- README_GGUF.md-about-gguf end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GGUF)
* [Jon Durbin's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/jondurbin/airoboros-l2-13b-3.0)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: Llama-2-Uncensored
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
```
<!-- prompt-template end -->
<!-- compatibility_gguf start -->
## Compatibility
These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221)
They are also compatible with many third party UIs and libraries - please see the list at the top of this README.
## Explanation of quantisation methods
<details>
<summary>Click to see details</summary>
The new methods available are:
* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
Refer to the Provided Files table below to see what files use which methods, and how.
</details>
<!-- compatibility_gguf end -->
<!-- README_GGUF.md-provided-files start -->
## Provided files
| Name | Quant method | Bits | Size | Max RAM required | Use case |
| ---- | ---- | ---- | ---- | ---- | ----- |
| [airoboros-l2-13b-3.0.Q2_K.gguf](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GGUF/blob/main/airoboros-l2-13b-3.0.Q2_K.gguf) | Q2_K | 2 | 5.43 GB| 7.93 GB | smallest, significant quality loss - not recommended for most purposes |
| [airoboros-l2-13b-3.0.Q3_K_S.gguf](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GGUF/blob/main/airoboros-l2-13b-3.0.Q3_K_S.gguf) | Q3_K_S | 3 | 5.66 GB| 8.16 GB | very small, high quality loss |
| [airoboros-l2-13b-3.0.Q3_K_M.gguf](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GGUF/blob/main/airoboros-l2-13b-3.0.Q3_K_M.gguf) | Q3_K_M | 3 | 6.34 GB| 8.84 GB | very small, high quality loss |
| [airoboros-l2-13b-3.0.Q3_K_L.gguf](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GGUF/blob/main/airoboros-l2-13b-3.0.Q3_K_L.gguf) | Q3_K_L | 3 | 6.93 GB| 9.43 GB | small, substantial quality loss |
| [airoboros-l2-13b-3.0.Q4_0.gguf](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GGUF/blob/main/airoboros-l2-13b-3.0.Q4_0.gguf) | Q4_0 | 4 | 7.37 GB| 9.87 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
| [airoboros-l2-13b-3.0.Q4_K_S.gguf](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GGUF/blob/main/airoboros-l2-13b-3.0.Q4_K_S.gguf) | Q4_K_S | 4 | 7.41 GB| 9.91 GB | small, greater quality loss |
| [airoboros-l2-13b-3.0.Q4_K_M.gguf](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GGUF/blob/main/airoboros-l2-13b-3.0.Q4_K_M.gguf) | Q4_K_M | 4 | 7.87 GB| 10.37 GB | medium, balanced quality - recommended |
| [airoboros-l2-13b-3.0.Q5_0.gguf](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GGUF/blob/main/airoboros-l2-13b-3.0.Q5_0.gguf) | Q5_0 | 5 | 8.97 GB| 11.47 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
| [airoboros-l2-13b-3.0.Q5_K_S.gguf](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GGUF/blob/main/airoboros-l2-13b-3.0.Q5_K_S.gguf) | Q5_K_S | 5 | 8.97 GB| 11.47 GB | large, low quality loss - recommended |
| [airoboros-l2-13b-3.0.Q5_K_M.gguf](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GGUF/blob/main/airoboros-l2-13b-3.0.Q5_K_M.gguf) | Q5_K_M | 5 | 9.23 GB| 11.73 GB | large, very low quality loss - recommended |
| [airoboros-l2-13b-3.0.Q6_K.gguf](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GGUF/blob/main/airoboros-l2-13b-3.0.Q6_K.gguf) | Q6_K | 6 | 10.68 GB| 13.18 GB | very large, extremely low quality loss |
| [airoboros-l2-13b-3.0.Q8_0.gguf](https://huggingface.co/TheBloke/airoboros-l2-13B-3.0-GGUF/blob/main/airoboros-l2-13b-3.0.Q8_0.gguf) | Q8_0 | 8 | 13.83 GB| 16.33 GB | very large, extremely low quality loss - not recommended |
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
<!-- README_GGUF.md-provided-files end -->
<!-- README_GGUF.md-how-to-download start -->
## How to download GGUF files
**Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.
The following clients/libraries will automatically download models for you, providing a list of available models to choose from:
- LM Studio
- LoLLMS Web UI
- Faraday.dev
### In `text-generation-webui`
Under Download Model, you can enter the model repo: TheBloke/airoboros-l2-13B-3.0-GGUF and below it, a specific filename to download, such as: airoboros-l2-13b-3.0.Q4_K_M.gguf.
Then click Download.
### On the command line, including multiple files at once
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
Then you can download any individual model file to the current directory, at high speed, with a command like this:
```shell
huggingface-cli download TheBloke/airoboros-l2-13B-3.0-GGUF airoboros-l2-13b-3.0.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
<details>
<summary>More advanced huggingface-cli download usage</summary>
You can also download multiple files at once with a pattern:
```shell
huggingface-cli download TheBloke/airoboros-l2-13B-3.0-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'
```
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/airoboros-l2-13B-3.0-GGUF airoboros-l2-13b-3.0.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
</details>
<!-- README_GGUF.md-how-to-download end -->
<!-- README_GGUF.md-how-to-run start -->
## Example `llama.cpp` command
Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
```shell
./main -ngl 32 -m airoboros-l2-13b-3.0.Q4_K_M.gguf --color -c 4096 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "[INST] <<SYS>>\nYou are a help, unbiased, uncensored assistant.\n<</SYS>\n\n{prompt} [/INST]"
```
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
Change `-c 4096` to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically.
If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`
For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md)
## How to run in `text-generation-webui`
Further instructions here: [text-generation-webui/docs/llama.cpp.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/llama.cpp.md).
## How to run from Python code
You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries.
### How to load this model in Python code, using ctransformers
#### First install the package
Run one of the following commands, according to your system:
```shell
# Base ctransformers with no GPU acceleration
pip install ctransformers
# Or with CUDA GPU acceleration
pip install ctransformers[cuda]
# Or with AMD ROCm GPU acceleration (Linux only)
CT_HIPBLAS=1 pip install ctransformers --no-binary ctransformers
# Or with Metal GPU acceleration for macOS systems only
CT_METAL=1 pip install ctransformers --no-binary ctransformers
```
#### Simple ctransformers example code
```python
from ctransformers import AutoModelForCausalLM
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
llm = AutoModelForCausalLM.from_pretrained("TheBloke/airoboros-l2-13B-3.0-GGUF", model_file="airoboros-l2-13b-3.0.Q4_K_M.gguf", model_type="llama", gpu_layers=50)
print(llm("AI is going to"))
```
## How to use with LangChain
Here are guides on using llama-cpp-python and ctransformers with LangChain:
* [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)
* [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)
<!-- README_GGUF.md-how-to-run end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Pierre Kircher, Stanislav Ovsiannikov, Michael Levine, Eugene Pentland, Andrey, 준교 김, Randy H, Fred von Graf, Artur Olbinski, Caitlyn Gatomon, terasurfer, Jeff Scroggin, James Bentley, Vadim, Gabriel Puliatti, Harry Royden McLaughlin, Sean Connelly, Dan Guido, Edmond Seymore, Alicia Loh, subjectnull, AzureBlack, Manuel Alberto Morcote, Thomas Belote, Lone Striker, Chris Smitley, Vitor Caleffi, Johann-Peter Hartmann, Clay Pascal, biorpg, Brandon Frisco, sidney chen, transmissions 11, Pedro Madruga, jinyuan sun, Ajan Kanaga, Emad Mostaque, Trenton Dambrowitz, Jonathan Leane, Iucharbius, usrbinkat, vamX, George Stoitzev, Luke Pendergrass, theTransient, Olakabola, Swaroop Kallakuri, Cap'n Zoog, Brandon Phillips, Michael Dempsey, Nikolai Manek, danny, Matthew Berman, Gabriel Tamborski, alfie_i, Raymond Fosdick, Tom X Nguyen, Raven Klaugh, LangChain4j, Magnesian, Illia Dulskyi, David Ziegler, Mano Prime, Luis Javier Navarrete Lozano, Erik Bjäreholt, 阿明, Nathan Dryer, Alex, Rainer Wilmers, zynix, TL, Joseph William Delisle, John Villwock, Nathan LeClaire, Willem Michiel, Joguhyik, GodLy, OG, Alps Aficionado, Jeffrey Morgan, ReadyPlayerEmma, Tiffany J. Kim, Sebastain Graf, Spencer Kim, Michael Davis, webtim, Talal Aujan, knownsqashed, John Detwiler, Imad Khwaja, Deo Leter, Jerry Meng, Elijah Stavena, Rooh Singh, Pieter, SuperWojo, Alexandros Triantafyllidis, Stephen Murray, Ai Maven, ya boyyy, Enrico Ros, Ken Nordquist, Deep Realms, Nicholas, Spiking Neurons AB, Elle, Will Dee, Jack West, RoA, Luke @flexchar, Viktor Bowallius, Derek Yates, Subspace Studios, jjj, Toran Billups, Asp the Wyvern, Fen Risland, Ilya, NimbleBox.ai, Chadd, Nitin Borwankar, Emre, Mandus, Leonard Tan, Kalila, K, Trailburnt, S_X, Cory Kujawski
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
<!-- original-model-card start -->
# Original model card: Jon Durbin's Airoboros L2 13B 3.0
### Overview
Another experimental model, using mostly sythetic data generated by [airoboros](https://github.com/jondurbin/airoboros)
#### IMPORTANT NOTE - llama-2 chat format!!!
The airobors 3.0 models have been updated to use llama-2 chat format, rather than vicuna style user/assistant!
This is a breaking change, although most inference systems support llama-2 chat templating.
See prompt format below for all of the details.
#### Highlights
A model built on [airoboros-3.0 dataset](https://hf.co/datasets/jondurbin/airoboros-3.0), which is a continuation of [airoboros-2.2.1 dataset](https://hf.co/datasets/jondurbin/airoboros-2.2.1) with the following extras:
* MathJSON - this is not yet part of the code directly, but can be easily parsed and executed, as a way to perform deterministic calculations. LLMs suck at math, so this does problem formulation and let's the work be done by traditional operations.
* Human-generated RP multi-turn conversation - provided by an anon, this is a human created RP dataset, meant to assist with multi-turn coherency and chat experience.
This is a fairly general purpose model, but focuses heavily on instruction following, rather than casual chat/roleplay.
Huge thank you to the folks over at [a16z](https://a16z.com/) for sponsoring the costs associated with building models and associated tools!
### Prompt format
The prompt format is llama-2 chat.
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
```
For multi-turn, the prompt format is as follows:
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt 0} [/INST] {response 0} </s><s>[INST] {prompt 1} [/INST] {response 1} </s><s>...[INST] {prompt N} [/INST]
```
Here is a simple python script that can help generate the prompt:
```python
def get_chat_prompt(message, chat_history, system_prompt):
texts = [f'[INST] <<SYS>>\n{system_prompt}\n<</SYS>>\n\n']
do_strip = False
for user_input, response in chat_history:
user_input = user_input.strip() if do_strip else user_input
do_strip = True
texts.append(f'{user_input} [/INST] {response.strip()} </s><s>[INST] ')
message = message.strip() if do_strip else message
texts.append(f'{message} [/INST] ')
return ''.join(texts)
def get_prompt(item):
"""
`item` is expected to be in chatml format, e.g.
{
"id": 1,
"conversations": [
{
"role": "system",
"content": "You are a helpful, unbiased, uncensored assistant."
},
{
"role": "user",
"content": "Hi!"
},
{
"role": "assistant",
"content": "Hello, how are you today?"
},
{
"role": "user",
"content": "Just peachy, and you?"
}
]
}
"""
system_prompt = share_gpt_item['conversations'][0]['value']
offset = 1
if share_gpt_item['conversations'][0]['role'] != 'system':
system_prompt = 'You are an unbiased, uncensored, helpful assistant.'
offset = 0
chat_history = [
(share_gpt_item['conversations'][idx]['content'], share_gpt_item['conversations'][idx + 1]['content'])
for idx in range(offset, len(item['conversations']) - 1, 2)
]
message = item['conversations'][-1]['content']
return get_chat_prompt(message, chat_history, system_prompt)
```
### Fine tuning info
https://wandb.ai/jondurbin/airoboros-l2-13b-3.0/runs/97utxyk9?workspace=user-jondurbin
### Helpful usage tips
#### MathJSON
Ask mathematical questions, prefixed with "Create a MathJSON solution to the following:", and you should get output with a formula that can be executed by https://cortexjs.io/compute-engine/
I also made a (really quite hacky, incomplete, semi-functional) python version that works for most expressions, see [mathjson.py](https://github.com/jondurbin/airoboros/blob/mathjson/airoboros/mathjson.py)
__Use a very low temperature!__
Example:
```
[INST] <<SYS>>
You are an assistant with vast knowledge in all things math.
<</SYS>
Create a MathJSON solution to the following: Calculate the area of a circle with radius of 17.2456cm. Include your reasoning. [/INST]
```
Output:
```
The area (A) of a circle is given by the formula A = πr², where r is the radius of the circle. Since the radius is given in centimeters, we need to convert it to meters first. We know that 1 meter equals 100 centimeters. Therefore, we can represent the radius as: r = 17.2456 cm * 100/1.
Once we have the radius in meters, we can calculate the area of the circle using the formula A = πr².
Solution as MathJSON:
<mathjson>
[
"Multiply",
"Pi",
[
"Power",
17.2456,
2
]
]
</mathjson>
```
You can then validate the JSON between `<mathjson>` and `</mathjson>`, then pass the parsed JSON to compute-engine JS or the `evaluate` function in mathjson.py to calculate the response.
#### Context obedient question answering
By obedient, I mean the model was trained to ignore what it thinks it knows, and uses the context to answer the question. The model was also tuned to limit the values to the provided context as much as possible to reduce hallucinations.
The format for a closed-context prompt is as follows:
```
BEGININPUT
BEGINCONTEXT
[key0: value0]
[key1: value1]
... other metdata ...
ENDCONTEXT
[insert your text blocks here]
ENDINPUT
[add as many other blocks, in the exact same format]
BEGININSTRUCTION
[insert your instruction(s). The model was tuned with single questions, paragraph format, lists, etc.]
ENDINSTRUCTION
```
It's also helpful to add "Don't make up answers if you don't know." to your instruction block to make sure if the context is completely unrelated it doesn't make something up.
*The __only__ prompts that need this closed context formating are closed-context instructions. Normal questions/instructions do not!*
I know it's a bit verbose and annoying, but after much trial and error, using these explicit delimiters helps the model understand where to find the responses and how to associate specific sources with it.
- `BEGININPUT` - denotes a new input block
- `BEGINCONTEXT` - denotes the block of context (metadata key/value pairs) to associate with the current input block
- `ENDCONTEXT` - denotes the end of the metadata block for the current input
- [text] - Insert whatever text you want for the input block, as many paragraphs as can fit in the context.
- `ENDINPUT` - denotes the end of the current input block
- [repeat as many input blocks in this format as you want]
- `BEGININSTRUCTION` - denotes the start of the list (or one) instruction(s) to respond to for all of the input blocks above.
- [instruction(s)]
- `ENDINSTRUCTION` - denotes the end of instruction set
It sometimes works without `ENDINSTRUCTION`, but by explicitly including that in the prompt, the model better understands that all of the instructions in the block should be responded to.
__Use a very low temperature!__
Here's a trivial, but important example to prove the point:
```
BEGININPUT
BEGINCONTEXT
date: 2021-01-01
url: https://web.site/123
ENDCONTEXT
In a shocking turn of events, blueberries are now green, but will be sticking with the same name.
ENDINPUT
BEGININSTRUCTION
What color are bluberries? Source?
ENDINSTRUCTION
```
And the response:
```
Blueberries are now green.
Source:
date: 2021-01-01
url: https://web.site/123
```
#### Summarization
500 samples have been included from [this dataset](https://huggingface.co/datasets/mattpscott/airoboros-summarization), using the same format as contextual question answering, for example:
```
BEGININPUT
{text to summarize}
ENDINPUT
BEGININSTRUCTION
Summarize the input in around 130 words.
ENDINSTRUCTION
```
#### Getting longer responses
You can use a few techniques to get longer responses.
Detailed prompts, with explicit instruction for word count:
```
Please compose a narrative set in the heart of an ancient library, steeped in the scent of old parchment and ink. The protagonist should be a young scholar who is dedicated to studying the art of storytelling and its evolution throughout history. In her pursuit of knowledge, she stumbles upon a forgotten tome that seems to possess an unusual aura. This book has the ability to bring stories to life, literally manifesting characters and scenarios from within its pages into reality.
The main character must navigate through various epochs of storytelling - from oral traditions of tribal societies, through medieval minstrels' tales, to modern-day digital narratives - as they come alive around her. Each era presents its unique challenges and lessons about the power and impact of stories on human civilization.
One such character could be a sentient quill pen, who was once used by renowned authors of yesteryears and now holds their wisdom and experiences. It becomes her mentor, guiding her through this journey with witty remarks and insightful commentary.
Ensure that your tale encapsulates the thrill of adventure, the beauty of learning, and the profound connection between humans and their stories. All characters involved should be non-human entities. Feel free to explore creative liberties but maintain the mentioned elements.
Your response should be approximately 2300 words.
```
Or, a simpler example:
```
Please create a long, detailed story about a dragon in an old growth forest who, for some reason, begins speaking the words of the source code of linux.
```
There are a few examples of next chapter completion as well, e.g.:
```
Write the next chapter of a historical fiction novel set in Paris during the 20th century.
Here's a summary of the previous chapter:
In the vibrant city of Paris, amid the tumultuous changes of the 20th century, our protagonist Margot, an aspiring fashion designer, has just secured an apprenticeship at a prestigious couture house. She meets Lucien, a charming journalist who covers the fashion industry. Together they navigate the ever-changing world of fashion and society, uncovering secrets that reveal the intricate links between style, politics, and culture. As the chapter concludes, they decide to delve deeper into the hidden corners of the fashion world to unravel its mysteries.
Requirements for the next chapter:
1. Character Development of Margot and Lucien:
- Margot's Evolution: Unfold more about Margot's past, her dreams of revolutionizing fashion, and her struggle to establish herself in a male-dominated industry. Illustrate her growing expertise, innovative ideas, and increasing dependence on Lucien.
- Lucien's Complexity: Introduce uncertainties surrounding Lucien's background and real motives. Increase suspense by suggesting undisclosed information he possesses, while also highlighting his wit and perceptiveness.
2. Exploration of Paris and the Couture House:
- Paris: Elaborate their journey through the bustling streets of Paris, including encounters with iconic figures, social unrest, and relics from different eras of French history.
- The Couture House: Expand on the grandeur of the couture house they work in, filled with artistic masterpieces, intense competition, and cryptic notes hinting at a scandalous past.
3. Emergence of the Subplot: The Lost Collection:
- Discovery: Have Margot and Lucien stumble upon a secret vault containing a lost collection designed before World War II, raising new questions about the previous owner and the influence of war on fashion.
- Revelation: Capture their shock as they realize the designs were plagiarized, the potential repercussions, and the opportunities it presents for Margot's career.
- Twist: End with a twist that suggests there are other stolen collections across Paris, setting up their new mission.
Your response should be approximately 650 words.
```
#### Coding
You can ask for fairly complex coding instructions with multiple criteria, e.g.:
```
Create a python application with the following requirements:
- Asyncio FastAPI webserver
- ping endpoint that returns the current date in JSON format
- file upload endpoint, which calculates the file's sha256 checksum, and checks postgres to deduplicate
```
Or inline criteria:
```
Write a multi-threaded TCP server in C that accepts a "GET [key]" input and "SET [key] [value]" input, and uses a binary tree to get and store the input values.
```
You can also optionally add a single space and "PLAINFORMAT" at the end of your prompt to avoid backticks, explanations, etc. and just print the code, e.g.:
```
Write a websocket application in node.js. PLAINFORMAT
```
#### Agent/function calling
The dataset includes many examples of function/args generation based on input criteria. This is somewhat similar to the OpenAI function calling, but the output is either JSON or YAML.
Example prompt:
```
As an AI assistant, please select the most suitable function and parameters from the list of available functions below, based on the user's input. Provide your response in JSON format.
Input: I want to know how many times 'Python' is mentioned in my text file.
Available functions:
file_analytics:
description: This tool performs various operations on a text file.
params:
action: The operation we want to perform on the data, such as "count_occurrences", "find_line", etc.
filters:
keyword: The word or phrase we want to search for.
```
Response:
```json
{
"function": "file_analytics",
"params": {
"action": "count_occurrences",
"filters": {
"keyword": "Python"
}
}
}
```
#### Chain-of-thought
You can ask for several possible responses to a given problem, with a ranking and final answer selection.
Example prompt:
```
A group of people decide to play a round-robin tournament where every player plays every other player exactly once. If a total of 45 games were played, how many players participated in the tournament? Offer several potential answers, rank them according to mathematical logic, and pick the most feasible one.
```
Example response:
```
Solution 1:
In a round-robin tournament, each player plays every other player exactly once. This means that the total number of games played is equal to the combination of all players taken two at a time (since a game involves two players).
The formula for combinations is nC2 = n(n-1)/2, where n is the total number of players. We know that nC2 equals 45 (the total number of games), so we can set up the equation as follows:
n(n-1)/2 = 45
Solving this equation gives us n=10.
Final answer: There were 10 players in the tournament.
Solution 2:
Another way to approach this problem is by using the concept of triangular numbers. Triangular numbers are a sequence of numbers where each number can be represented as a triangle with dots. The nth triangular number is given by the sum of the first n natural numbers.
If we consider each game as a dot and each player as a side of the triangle, then the total number of games (45) would correspond to the 9th triangular number because 1+2+3+4+5+6+7+8+9=45. However, since each side of the triangle represents a player, and there's one more player than sides in our model (because the last player has no one left to play against), we need to add one more to get the total number of players.
Final answer: There were 10 players in the tournament.
Solution 3:
We could also use trial and error to solve this problem. Starting from a small number of players, we calculate how many games would be played in a round-robin tournament until we reach 45 games. For example, with 2 players, only 1 game is played; with 3 players, 3 games are played; with 4 players, 6 games are played, and so on. Continuing this process, we find that with 10 players, 45 games are played.
Final answer: There were 10 players in the tournament.
Ranking of solutions from best to worst:
1. Solution 1: This solution uses direct mathematical reasoning based on the properties of a round-robin tournament and solves an equation to find the answer.
2. Solution 2: This solution uses the concept of triangular numbers which requires additional steps and knowledge beyond the immediate problem context, but still arrives at the correct answer.
3. Solution 3: This solution relies on trial and error, which is less efficient and reliable than the other methods, but it eventually gets to the right answer.
Best and final answer: There were 10 players in the tournament.
```
#### reWOO style execution planning
The model now supports execution planning for complex instructions that would require making use of several tools. The output is just the plan, you must implement a mechanism to parse the output and actually call the functions!
Example prompt:
```
Please construct a systematic plan to generate an optimal response to the user instruction, utilizing a set of provided tools. Each plan will correspond to an evidence value, which will be the output of one of the available functions given an input string
that could be the user's question, one or more prior evidence values, or a combination of both.
Here are the tools available to be called:
DuckDuckGo[input]: Tool that enables the user to search for information using the DuckDuckGo search engine. This tool is helpful in locating a suitable list of sites that may contain the answer to the user's query. It does not directly answer the question but finds a list of sites that might have the answer.
TextScraper[input]: Load one or more websites from the input string containing newline delimited links, where input is one or more links, and produces plain text output containing the content of the links.
HyperlinkExtractor[input]: Extract links from plain text and produces a plain text, newline delimited response of links.
KnowledgeModel[input]: Question answering language model, particularly useful in answering questions based on an input passage of text. The input must be a text question that references an :evidence[n]: variable, e.g. What color is the cat, given :evidence1:?
The output should be in format:
Plan: [first action to take based on input question]
:evidence0: = [function to call with input parameter]
Plan: [next action to take, based on result of :evidence0:]
:evidence1: = [next function to call and input parameter, which may include reference to previous evidence, e.g. "Given context :evidence0"]
...
Answer: [:evidence[n]: containing the final answer.]
Question: Who won the Nobel Prize in Literature in 2020 and what is their most famous work?
```
Response:
```
Plan: Begin by conducting a web search to find out who won the Nobel Prize in Literature in 2020.
:evidence0: = DuckDuckGo[Nobel Prize in Literature 2020 winner]
Plan: Utilize the KnowledgeModel to interpret the search results and find the name of the laureate.
:evidence1: = KnowledgeModel[Who won the Nobel Prize in Literature in 2020, given :evidence0:?]
Plan: Conduct another web search to find the most famous work of the identified laureate.
:evidence2: = DuckDuckGo[Most famous work of :evidence1:]
Plan: Extract the relevant links from the DuckDuckGo search results for a more focused search.
:evidence3: = HyperlinkExtractor[:evidence2:]
Plan: Use the TextScraper tool to extract information from the relevant links.
:evidence4: = TextScraper[:evidence3:]
Plan: Finally, utilize the KnowledgeModel to identify and summarize the most famous work of the laureate from the extracted information.
:evidence5: = KnowledgeModel[What is the most famous work of :evidence1:, given :evidence4:?]
Answer: :evidence5:
```
For this to be useful, you'd have to parse the output plan text, and implement/call each of the functions. This is just pseudo-code, completely untested off the top of my head, and obviously would requiring full implementation + hardening:
```python
import re
import requests
def inject_context(input_text, **context):
for ref in set(re.findall(r"(:evidence[0-9]+:)", input_text, re.I)):
input_text = input_text.replace(ref, context.get(ref, ""))
return input_text
def duckduckgo(input_text, **context):
search_string = inject_context(input_text, **context)
... search via duck duck go using search_string
... return text content
def link_extractor(input_text, **context):
input_text = inject_context(input_text, **context)
return "\n".join(list(set(re.findall(r"(https?://[^\s]+?\.?)", input_text, re.I))))
def scrape(input_text, **context):
input_text = inject_context(input_text, **context)
text = []
for link in input_text.splitlines():
text.append(requests.get(link).text)
return "\n".join(text)
def infer(input_text, **context)
prompt = inject_context(input_text, **context)
... call model with prompt, return output
def parse_plan(plan):
method_map = {
"DuckDuckGo": duckduckgo,
"HyperlinkExtractor": link_extractor,
"KnowledgeModel": infer,
"TextScraper": scrape,
}
context = {}
for line in plan.strip().splitlines():
if line.startswith("Plan:"):
print(line)
continue
parts = re.match("^(:evidence[0-9]+:)\s*=\s*([^\[]+])(\[.*\])\s$", line, re.I)
if not parts:
if line.startswith("Answer: "):
return context.get(line.split(" ")[-1].strip(), "Answer couldn't be generated...")
raise RuntimeError("bad format: " + line)
context[parts.group(1)] = method_map[parts.group(2)](parts.group(3), **context)
```
### Contribute
If you're interested in new functionality, particularly a new "instructor" type to generate a specific type of training data,
take a look at the dataset generation tool repo: https://github.com/jondurbin/airoboros and either make a PR or open an issue with details.
To help me with the OpenAI/compute costs:
- https://bmc.link/jondurbin
- ETH 0xce914eAFC2fe52FdceE59565Dd92c06f776fcb11
- BTC bc1qdwuth4vlg8x37ggntlxu5cjfwgmdy5zaa7pswf
### Licence and usage restrictions
The airoboros 3.0 models are built on top of multiple base models, each with their own license/restrictions.
The models with `-l2` in the name have a custom Meta license:
- See the [meta-license/LICENSE.txt](meta-license/LICENSE.txt) file attached for the original license provided by Meta.
- See also [meta-license/USE_POLICY.md](meta-license/USE_POLICY.md) and [meta-license/Responsible-Use-Guide.pdf](meta-license/Responsible-Use-Guide.pdf), also provided by Meta.
The models with `-m-` are mistral-7b (apache 2.0)
The model with `-3b` uses Stability AI, which as a `cc-by-sa-4.0` license.
The fine-tuning data was mostly generated by OpenAI API calls to gpt-4, via [airoboros](https://github.com/jondurbin/airoboros)
The ToS for OpenAI API usage has a clause preventing the output from being used to train a model that __competes__ with OpenAI
- what does *compete* actually mean here?
- these small open source models will not produce output anywhere near the quality of gpt-4, or even gpt-3.5, so I can't imagine this could credibly be considered competing in the first place
- if someone else uses the dataset to do the same, they wouldn't necessarily be violating the ToS because they didn't call the API, so I don't know how that works
- the training data used in essentially all large language models includes a significant amount of copyrighted or otherwise non-permissive licensing in the first place
- other work using the self-instruct method, e.g. the original here: https://github.com/yizhongw/self-instruct released the data and model as apache-2
I am purposingly leaving this license ambiguous (other than the fact you must comply with the Meta original license for llama-2) because I am not a lawyer and refuse to attempt to interpret all of the terms accordingly.
Your best bet is probably to avoid using this commercially due to the OpenAI API usage.
Either way, by using this model, you agree to completely indemnify me.
<!-- original-model-card end -->
|
adamo1139/PS_AD_O365_Mistral_superCOT_7B_03_QLoRA_GGUF
|
adamo1139
| 2023-10-05T22:44:49Z | 3 | 0 | null |
[
"gguf",
"dataset:adamo1139/PS_AD_Office365_03",
"endpoints_compatible",
"region:us"
] | null | 2023-10-05T20:53:34Z |
---
datasets:
- adamo1139/PS_AD_Office365_03
---
quantization of the PS_AD_O365 QLoRA with mistral 7b superCOT.
This version was trained on a big subset of coding dataset from airoboros. 1 epoch.
You can ask it questions about PowerShell, AD or Office 365. I am not sure if it's better than mistral-instruct, consider this moreso an experiment rather than a tool.
Prompt template:
A collaborative discussion between Pupil and Teacher.Pupil:{question}Teacher{answer}
|
Johnyquest7/thyroid_open_llama_3b_v2
|
Johnyquest7
| 2023-10-05T22:36:35Z | 1 | 0 |
peft
|
[
"peft",
"region:us"
] | null | 2023-10-05T22:26:05Z |
---
library_name: 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: float16
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: float16
### Framework versions
- PEFT 0.5.0
- PEFT 0.5.0
|
alikanakar/xlsr-syntesized-turkish-2-hour-llr
|
alikanakar
| 2023-10-05T21:50:45Z | 3 | 0 |
transformers
|
[
"transformers",
"pytorch",
"tensorboard",
"wav2vec2",
"automatic-speech-recognition",
"generated_from_trainer",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] |
automatic-speech-recognition
| 2023-10-05T20:07:31Z |
---
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: xlsr-syntesized-turkish-2-hour-llr
results: []
---
<!-- 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. -->
# xlsr-syntesized-turkish-2-hour-llr
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 3.1105
- Wer: 1.0
## 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: 1e-05
- train_batch_size: 2
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 200
- num_epochs: 20
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:-----:|:----:|:---------------:|:---:|
| 9.4081 | 1.04 | 100 | 8.3205 | 1.0 |
| 5.0295 | 2.08 | 200 | 4.3245 | 1.0 |
| 3.7196 | 3.12 | 300 | 3.8079 | 1.0 |
| 3.4558 | 4.17 | 400 | 3.5232 | 1.0 |
| 3.304 | 5.21 | 500 | 3.3519 | 1.0 |
| 3.2115 | 6.25 | 600 | 3.2609 | 1.0 |
| 3.1533 | 7.29 | 700 | 3.1903 | 1.0 |
| 3.0813 | 8.33 | 800 | 3.1409 | 1.0 |
| 3.0279 | 9.38 | 900 | 3.0815 | 1.0 |
| 2.9087 | 10.42 | 1000 | 3.0615 | 1.0 |
| 2.6893 | 11.46 | 1100 | 3.0408 | 1.0 |
| 2.4277 | 12.5 | 1200 | 3.0961 | 1.0 |
| 2.2325 | 13.54 | 1300 | 3.0973 | 1.0 |
| 2.0858 | 14.58 | 1400 | 3.1216 | 1.0 |
| 1.9604 | 15.62 | 1500 | 3.1271 | 1.0 |
| 1.8766 | 16.67 | 1600 | 3.1099 | 1.0 |
| 1.8093 | 17.71 | 1700 | 3.1191 | 1.0 |
| 1.7577 | 18.75 | 1800 | 3.1191 | 1.0 |
| 1.7174 | 19.79 | 1900 | 3.1105 | 1.0 |
### Framework versions
- Transformers 4.26.0
- Pytorch 2.0.1+cu118
- Datasets 2.9.0
- Tokenizers 0.13.3
|
Subsets and Splits
Filtered Qwen2.5 Distill Models
Identifies specific configurations of models by filtering cards that contain 'distill', 'qwen2.5', '7b' while excluding certain base models and incorrect model ID patterns, uncovering unique model variants.
Filtered Model Cards Count
Finds the count of entries with specific card details that include 'distill', 'qwen2.5', '7b' but exclude certain base models, revealing valuable insights about the dataset's content distribution.
Filtered Distill Qwen 7B Models
Filters for specific card entries containing 'distill', 'qwen', and '7b', excluding certain strings and patterns, to identify relevant model configurations.
Filtered Qwen-7b Model Cards
The query performs a detailed filtering based on specific keywords and excludes certain entries, which could be useful for identifying a specific subset of cards but does not provide deeper insights or trends.
Filtered Qwen 7B Model Cards
The query filters for specific terms related to "distilled" or "distill", "qwen", and "7b" in the 'card' column but excludes certain base models, providing a limited set of entries for further inspection.
Qwen 7B Distilled Models
The query provides a basic filtering of records to find specific card names that include keywords related to distilled Qwen 7b models, excluding a particular base model, which gives limited insight but helps in focusing on relevant entries.
Qwen 7B Distilled Model Cards
The query filters data based on specific keywords in the modelId and card fields, providing limited insight primarily useful for locating specific entries rather than revealing broad patterns or trends.
Qwen 7B Distilled Models
Finds all entries containing the terms 'distilled', 'qwen', and '7b' in a case-insensitive manner, providing a filtered set of records but without deeper analysis.
Distilled Qwen 7B Models
The query filters for specific model IDs containing 'distilled', 'qwen', and '7b', providing a basic retrieval of relevant entries but without deeper analysis or insight.
Filtered Model Cards with Distill Qwen2.
Filters and retrieves records containing specific keywords in the card description while excluding certain phrases, providing a basic count of relevant entries.
Filtered Model Cards with Distill Qwen 7
The query filters specific variations of card descriptions containing 'distill', 'qwen', and '7b' while excluding a particular base model, providing limited but specific data retrieval.
Distill Qwen 7B Model Cards
The query filters and retrieves rows where the 'card' column contains specific keywords ('distill', 'qwen', and '7b'), providing a basic filter result that can help in identifying specific entries.