The production environment deployment can be implemented in three phases:
- environmental preparation::
- Clone the GitHub repository and install the dependencies:
git clone https://github.com/TencentARC/ARC-Hunyuan-Video-7B && pip install -r requirements.txt - Download the 7B weights file from Hugging Face to the
experiments/pretrained_models/Catalog.
- Clone the GitHub repository and install the dependencies:
- Acceleration configuration: Install the vLLM library (
pip install vllm), and throughtest_setup.pyVerify CUDA compatibility. - service encapsulation::
- To integrate the core functionality (e.g., the
inference.py) encapsulated as a REST API, it is recommended to use the FastAPI framework. - Realize automatic segmentation processing pipeline for long videos.
- Ensure environment consistency with Docker containerized deployment.
- To integrate the core functionality (e.g., the
Following this process, the base deployment can be completed in less than 2 hours, with actual throughput dependent on GPU configuration.
This answer comes from the articleARC-Hunyuan-Video-7B: An Intelligent Model for Understanding Short Video ContentThe
May not be reproduced without permission:AI productivity tools " How to quickly deploy ARC-Hunyuan-Video-7B to production environment?
English 
简体中文 
日本語 
Deutsch 
Português do Brasil