xiejiahao/gemma4-bi150
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

Unified Gemma 4 family image for Gemma-4-31B and GEV-26B-Decide
All checks were successful
Docker Build and Push / docker (push) Successful in 1m11s
Details

Browse Source 
- detect_model.py picks TP/port/served name/eager/util from config.json
- validate TP against the model config (dense global KV heads, MoE k % 32)
- one apply_patch.py covers both variants, idempotent
- Dockerfile stays generic: no ENV/WORKDIR/EXPOSE/ENTRYPOINT
- run.sh/verify.sh follow the llama4-bi150 layout
This commit is contained in:
xiejiahao
2026-10-27 14:19:51 +08:00
commit 05f1c3932f
12 changed files with 1652 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
.gitea/workflows/task_info.env Normal file
View File

@@ -0,0 +1,3 @@
FRAMEWORK=vllm_gemma_4
GPU_TYPE=Iluvatar_bi-150
TASK_TYPE=text-generation