LLM in GitHub Actions
Version updated for https://github.com/muhac/llm-actions to version v2.0.0.
- This action is used across all versions by 2 repositories.
Action Type
This is a Composite action.
Go to the GitHub Marketplace to find the latest changes.
Action Summary
This GitHub Action allows users to run Hugging Face language models on a temporary vLLM server within a CI pipeline. It automates the process of deploying and performing inference with large language models directly in GitHub Actions, addressing tasks such as testing and model evaluation without requiring external infrastructure. Key capabilities include support for various model versions, handling private/public models, and optimizing resource usage within CI environments.
Release notes
What’s Changed
- Switch vLLM deployment from source build to Docker by @muhac in https://github.com/muhac/llm-actions/pull/5
- Require model name to run workflow by @Copilot in https://github.com/muhac/llm-actions/pull/4
Full Changelog: https://github.com/muhac/llm-actions/compare/v1.0.3...v2.0.0