MovieChat: From Dense Token to Sparse Memory for Long Video Understanding
Paper • 2307.16449 • Published • 17
How to use from
SGLangUse 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 "lmms-lab/MovieChat-ckpt" \
--host 0.0.0.0 \
--port 30000# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
-H "Content-Type: application/json" \
--data '{
"model": "lmms-lab/MovieChat-ckpt",
"prompt": "Once upon a time,",
"max_tokens": 512,
"temperature": 0.5
}'Quick Links
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
checkpoints for
@article{song2023moviechat,
title={MovieChat: From Dense Token to Sparse Memory for Long Video Understanding},
author={Song, Enxin and Chai, Wenhao and Wang, Guanhong and Zhang, Yucheng and Zhou, Haoyang and Wu, Feiyang and Guo, Xun and Ye, Tian and Lu, Yan and Hwang, Jenq-Neng and others},
journal={arXiv preprint arXiv:2307.16449},
year={2023}
}
- 7B is the vicuna v0 ckpt after applying delta on llama.
- Others are the ckpts acquire from the link on GitHub.
Prepared checkpoints here so that people can better acquire their models
- Downloads last month
- 22
Install from pip and serve model
# Install SGLang from pip: pip install sglang# Start the SGLang server: python3 -m sglang.launch_server \ --model-path "lmms-lab/MovieChat-ckpt" \ --host 0.0.0.0 \ --port 30000# Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "lmms-lab/MovieChat-ckpt", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'