Skip to content

DCC-BS/vllm

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

11 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

vLLM Patched Image

This repository provides a custom build of vLLM that includes a patched version of the transformers library installed directly from source.

Intent

The primary purpose is to maintain a Docker image compatible with vllm/vllm-openai but with the latest transformers features or fixes from the huggingface/transformers main branch. This is useful when the served model requires a specific transformer version that hasn't been officially released or integrated into the upstream vLLM images yet. An example of such a model is GLM-OCR.

How it works

The Dockerfile uses an ARG BASE_TAG to pull a specific version of vllm/vllm-openai and then updates the transformers package using uv.

CI/CD Workflow

The GitHub Actions workflow Build and Publish Patched vLLM can be triggered manually via workflow_dispatch.

Inputs

  • vllm_tag: The upstream vLLM image tag to use as a base (e.g., v0.7.2-cu124). Defaults to latest.

Outputs

The workflow builds and pushes the image to GitHub Container Registry (GHCR):

  • ghcr.io/DCC-BS/vllm:<vllm_tag>
  • ghcr.io/DCC-BS/vllm:latest

Usage

You can pull the latest patched image:

docker pull ghcr.io/DCC-BS/vllm:latest

Or a specific version:

docker pull ghcr.io/DCC-BS/vllm:v0.16.0-cu130

About

Patched vLLM docker image with updated transformers version (>5)

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors