Update Dockerfile.ci to install vLLM from a new commit and simplify worker_type retrieval in omni_stage.py. Removed unnecessary dependency installations for flashinfer and numpy.

Signed-off-by: tzhouam <tzhouam@connect.ust.hk>

docker/Dockerfile.ci

@@ -13,20 +13,12 @@ RUN apt-get update && \
 
 # Install vLLM wheel for the specific commit
 # URL format: https://wheels.vllm.ai/<commit_hash>/
-# Current vLLM wheel commit: 6521ccf2860eb72cf23a9fa9044bf9721fd9a7c6
+# Current vLLM wheel commit: 097eb544e9a22810c9b7a59e586b61627b308362
 # Note: --prerelease=allow is required because nightly/commit wheels use
 # pre-release version numbers (e.g. 2.10.0.dev...) which uv ignores by default,
 # causing it to fall back to the older stable release from PyPI.
-RUN uv pip install --system --upgrade --force-reinstall --prerelease=allow vllm --extra-index-url https://wheels.vllm.ai/6521ccf2860eb72cf23a9fa9044bf9721fd9a7c6
+RUN uv pip install --system --upgrade --force-reinstall --prerelease=allow vllm --extra-index-url https://wheels.vllm.ai/097eb544e9a22810c9b7a59e586b61627b308362
 
 RUN uv pip install --system --no-cache-dir ".[dev]"
-RUN uv pip install --system --upgrade \
-    "flashinfer-cubin==0.6.4" \
-    "nvidia-cublas-cu12==12.9.1.4" \
-    "numpy==2.2.6"
-RUN uv pip install --system --upgrade \
-    "flashinfer-jit-cache==0.6.4" \
-    --index-url https://flashinfer.ai/whl/cu129
-RUN ln -sf /usr/bin/python3 /usr/bin/python
 
 ENTRYPOINT []

vllm_omni/entrypoints/omni_stage.py

@@ -210,7 +210,7 @@ def _sequential_init_lock(engine_args: dict[str, Any], stage_init_timeout: int =
 
 
 def _resolve_worker_cls(engine_args: dict[str, Any]) -> None:
-    worker_type = engine_args.pop("worker_type", None)
+    worker_type = engine_args.get("worker_type", None)
     if not worker_type:
         return
     worker_cls = engine_args.get("worker_cls")