Update README.md

README.md

@@ -1,199 +1,132 @@
 ---
-library_name: transformers
-tags: []
+license: mit
+pipeline_tag: image-feature-extraction
+tags:
+- pretrained
+datasets:
+- ylecun/mnist
 ---
 
-# Model Card for Model ID
+# Model Card for Llava-mnist
 
-<!-- Provide a quick summary of what the model is/does. -->
+This model is a simple linear layer vision encoder trained on the MNIST dataset, following the Llava training approach.
 
-
-
-## Model Details
-
-### Model Description
-
-<!-- Provide a longer summary of what this model is. -->
-
-This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
-
-- **Developed by:** [More Information Needed]
-- **Funded by [optional]:** [More Information Needed]
-- **Shared by [optional]:** [More Information Needed]
-- **Model type:** [More Information Needed]
-- **Language(s) (NLP):** [More Information Needed]
-- **License:** [More Information Needed]
-- **Finetuned from model [optional]:** [More Information Needed]
-
-### Model Sources [optional]
-
-<!-- Provide the basic links for the model. -->
-
-- **Repository:** [More Information Needed]
-- **Paper [optional]:** [More Information Needed]
-- **Demo [optional]:** [More Information Needed]
-
-## Uses
-
-<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
-
-### Direct Use
-
-<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
-
-[More Information Needed]
-
-### Downstream Use [optional]
-
-<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
-
-[More Information Needed]
-
-### Out-of-Scope Use
-
-<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
-
-[More Information Needed]
-
-## Bias, Risks, and Limitations
-
-<!-- This section is meant to convey both technical and sociotechnical limitations. -->
-
-[More Information Needed]
-
-### Recommendations
-
-<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
-
-Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
-
-## How to Get Started with the Model
-
-Use the code below to get started with the model.
-
-[More Information Needed]
+You can use this model alongside [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct).
 
 ## Training Details
+The model was trained on the *chat-style* MNIST dataset, which is structured as follows:
 
-### Training Data
-
-<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
-
-[More Information Needed]
-
-### Training Procedure
-
-<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
-
-#### Preprocessing [optional]
-
-[More Information Needed]
-
-
-#### Training Hyperparameters
-
-- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
-
-#### Speeds, Sizes, Times [optional]
-
-<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
-
-[More Information Needed]
-
-## Evaluation
-
-<!-- This section describes the evaluation protocols and provides the results. -->
-
-### Testing Data, Factors & Metrics
-
-#### Testing Data
-
-<!-- This should link to a Dataset Card if possible. -->
-
-[More Information Needed]
-
-#### Factors
-
-<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
-
-[More Information Needed]
-
-#### Metrics
-
-<!-- These are the evaluation metrics being used, ideally with a description of why. -->
-
-[More Information Needed]
-
-### Results
-
-[More Information Needed]
-
-#### Summary
-
-
-
-## Model Examination [optional]
-
-<!-- Relevant interpretability work for the model goes here -->
-
-[More Information Needed]
-
-## Environmental Impact
-
-<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+prompt: “\<image\>What digit is this?”
 
-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).
+output: "The digit is {label}."
 
-- **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]
+The Llava-MNIST model transforms the digit image into an embedding vector that resides in the same space as the text token embedding.
 
-## Technical Specifications [optional]
+The loss function optimized during training is defined as:
 
-### Model Architecture and Objective
+$L(W)= -\log P_W(This digit is \{label\}|\<image\>What digit is this?)$
 
-[More Information Needed]
+During training, the parameters of the Llama 3.1 model are kept frozen, and only the parameters of the vision encoder (Llava-MNIST) are optimized.
 
-### Compute Infrastructure
+## How to use
 
-[More Information Needed]
+You can input multi-modal data (vision and text) into the Llama 3.1 model by using the Llava-MNIST model as the vision encoder.
 
-#### Hardware
+```
+from transformers import AutoTokenizer, AutoModelForCausalLM
+import torch
+from datasets import load_dataset
+from torchvision import transforms
+import util
+from transformers import AutoModel
 
-[More Information Needed]
 
-#### Software
+def build_multi_modal_prompt(
+    prompt: str,
+    image: torch.Tensor,
+    tokenizer: AutoTokenizer,
+    model: AutoModelForCausalLM,
+    vision_model: AutoModel,
+) -> torch.Tensor:
+    parts = prompt.split("<image>")
+    prefix = tokenizer(parts[0])
+    suffix = tokenizer(parts[1])
+    prefix_embedding = model.get_input_embeddings()(torch.tensor(prefix["input_ids"]))
+    suffix_embedding = model.get_input_embeddings()(torch.tensor(suffix["input_ids"]))
+    image_embedding = vision_model(image).to(torch.bfloat16).to(model.device)
+    multi_modal_embedding = torch.cat(
+        [prefix_embedding, image_embedding, suffix_embedding], dim=0
+    )
+    return multi_modal_embedding
 
-[More Information Needed]
 
-## Citation [optional]
+model_id = "meta-llama/Meta-Llama-3.1-8B-Instruct"
 
-<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+model = AutoModelForCausalLM.from_pretrained(
+    model_id,
+    torch_dtype=torch.bfloat16,
+    device_map="auto",
+)
 
-**BibTeX:**
+vision_model = AutoModel.from_pretrained(
+    "speed/llava-mnist", trust_remote_code=True
+)
 
-[More Information Needed]
+terminators = [
+    tokenizer.eos_token_id,
+    tokenizer.convert_tokens_to_ids("<|eot_id|>"),
+]
 
-**APA:**
+system_prompt = (
+    "<|begin_of_text|><|start_header_id|>system<|end_header_id|><|eot_id|>"
+)
+user_prompt = "<|start_header_id|>user<|end_header_id|>"
+question = "<image>What digit is this?"
+assistant_prompt = "<|start_header_id|>assistant<|end_header_id|>"
 
-[More Information Needed]
+prompt = system_prompt + user_prompt + question + assistant_prompt
 
-## Glossary [optional]
+ds = load_dataset("ylecun/mnist", split="test")
 
-<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
 
-[More Information Needed]
+def transform_image(examples):
+    transform = transforms.Compose(
+        [
+            transforms.ToTensor(),
+            transforms.Normalize((0.1307,), (0.3081,)),
+            transforms.Lambda(lambda x: torch.flatten(x)),
+        ]
+    )
+    examples["pixel_values"] = [transform(image) for image in examples["image"]]
 
-## More Information [optional]
+    return examples
 
-[More Information Needed]
+ds.set_transform(transform = transform_image)
 
-## Model Card Authors [optional]
 
-[More Information Needed]
+model.eval()
+vision_model.eval()
 
-## Model Card Contact
+example = ds[0]
 
-[More Information Needed]
+input_embeded = util.build_multi_modal_prompt(
+    prompt, example["pixel_values"].unsqueeze(0), tokenizer, model, vision_model
+).unsqueeze(0)
+response = model.generate(
+    inputs_embeds=input_embeded,
+    max_new_tokens=20,
+    eos_token_id=terminators,
+    do_sample=True,
+    temperature=0.6,
+    top_p=0.9,
+)
+response = response[0]
+print("Label:", example["label"]) # Label: 7
+answer = tokenizer.decode(response, skip_special_tokens=True)
+print("Answer:", answer) # Answer: The digit is 7.
+
+```
+
+## References
+- Liu et al., LLaVA: Large Language and Vision Assistant, https://llava-vl.github.io/