Instructions to use MYTH-Lab/VW-LMM-Mistral-7b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use MYTH-Lab/VW-LMM-Mistral-7b with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="MYTH-Lab/VW-LMM-Mistral-7b")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForCausalLM

processor = AutoProcessor.from_pretrained("MYTH-Lab/VW-LMM-Mistral-7b")
model = AutoModelForCausalLM.from_pretrained("MYTH-Lab/VW-LMM-Mistral-7b")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = processor.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use MYTH-Lab/VW-LMM-Mistral-7b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "MYTH-Lab/VW-LMM-Mistral-7b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MYTH-Lab/VW-LMM-Mistral-7b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/MYTH-Lab/VW-LMM-Mistral-7b

SGLang

How to use MYTH-Lab/VW-LMM-Mistral-7b with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "MYTH-Lab/VW-LMM-Mistral-7b" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MYTH-Lab/VW-LMM-Mistral-7b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "MYTH-Lab/VW-LMM-Mistral-7b" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MYTH-Lab/VW-LMM-Mistral-7b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use MYTH-Lab/VW-LMM-Mistral-7b with Docker Model Runner:
```
docker model run hf.co/MYTH-Lab/VW-LMM-Mistral-7b
```

VW-LMM Model Card

This repo contains the weights of VW-LMM-Mistral-7b proposed in paper "Multi-modal Auto-regressive Modeling via Visual Words"

For specific usage and chat templates, please refer to our project repo https://github.com/pengts/VW-LMM

Model details

Model type: VW-LMM is an open-source chatbot trained by fine-tuning LLM on multimodal instruction-following data. It is an auto-regressive language model, based on the transformer architecture. Base LLM: mistralai/Mistral-7B-Instruct-v0.2

paper: https://arxiv.org/abs/2403.07720

code: https://github.com/pengts/VW-LMM

License

mistralai/Mistral-7B-Instruct-v0.2 license.

Citation

If you find our paper and code useful in your research, please consider giving a star :star: and citation :pencil:.

@misc{peng2024multimodal,
      title={Multi-modal Auto-regressive Modeling via Visual Words}, 
      author={Tianshuo Peng and Zuchao Li and Lefei Zhang and Hai Zhao and Ping Wang and Bo Du},
      year={2024},
      eprint={2403.07720},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}

Downloads last month: 26

Safetensors

Model size

8B params

Tensor type

BF16

Paper for MYTH-Lab/VW-LMM-Mistral-7b

Multi-modal Auto-regressive Modeling via Visual Words

Paper • 2403.07720 • Published Mar 12, 2024