zhoushasha/step3p5-vllm-nv
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

Add NVIDIA vLLM image and ModelHub release workflow
Some checks failed
Docker Build and Push / docker (push) Failing after 2s
Details

Browse Source
This commit is contained in:
zhousha
2026-09-29 15:39:44 +08:00
commit 7a7c4986b9
5 changed files with 164 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
Dockerfile Normal file
View File

@@ -0,0 +1,5 @@
FROM harbor.4pd.io/hardcore-tech/vllm/vllm-openai:v0.25.0-cu129
# Text inference does not require the incompatible TorchCodec video decoder.
RUN python3 -m pip uninstall -y torchcodec
RUN python3 -c "import vllm.entrypoints.openai.api_server; print('API server import OK')"