Skip to content

Conversation

@BaoZhuhan
Copy link

Summary

This PR adds comprehensive multi-environment support to Nano-vLLM, enabling easier deployment and reproducibility across different platforms. The changes include pip, conda, and Docker installation methods while maintaining the existing core functionality.

Changes Made

  • requirements.txt: Added pip requirements file with all necessary dependencies for easy pip installation
  • environment.yml: Created conda environment configuration with Python 3.12, PyTorch, CUDA support, and optimized channels (including Tsinghua mirrors for faster downloads)
  • Dockerfile: Provided containerized deployment with NVIDIA GPU support, volume mapping, and example execution
  • README.md: Updated installation section with detailed instructions for all three deployment methods (pip, conda, Docker)

Testing

  • Verified Python version compatibility (requires Python >=3.10, <3.13)
  • Environment files generated based on pyproject.toml dependencies
  • Documentation updated with clear installation steps

Motivation

Enhances project accessibility by providing multiple deployment options, making it easier for users to set up and reproduce the Nano-vLLM environment locally or in production.

Add requirements.txt for pip installation
Add environment.yml for conda environment setup
Add Dockerfile for containerized deployment with GPU support
Update README.md with installation instructions for all three methods
Create Engineer.csv to track changes
Add .gitignore in prompt/ folder to exclude it from commits
This enhances deployment flexibility while maintaining core functionality.
- Add requirements.txt for pip installation
- Add environment.yml for conda environment setup
- Add Dockerfile for containerized deployment with GPU support
- Update README.md with installation instructions for all three methods
This reverts commit 762c319, reversing
changes made to 1711884.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant