Fix Qwen3Omni VisionTransformer compatibility with updated vLLM Qwen2_5_VisionAttention

The upstream vLLM Qwen2_5_VisionAttention.forward() now requires a
`sequence_lengths` argument, but Qwen3_VisionBlock.forward() does not
pass it, causing a TypeError at init time.

Instead of patching the vLLM package directly, introduce a
Qwen3Omni_VisionTransformer subclass in vllm-omni that overrides
forward() to properly compute and propagate sequence_lengths via
MMEncoderAttention (following qwen3_vl.py's pattern). Both thinker
and talker now import the fixed subclass.

Also bump the vLLM wheel to e196028dbaa41b.

Made-with: Cursor

Signed-off-by: tzhouam <tzhouam@connect.ust.hk>
This commit is contained in:
tzhouam committed 2026-03-04 02:51:02 +00:00
1 parent 4914466f0b
commit ae818f4b45
3 files changed
+93 -4

No files matched your search

+1 -1
View File
@@ -18,7 +18,7 @@ RUN apt-get update && \
# 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/3dd142adb7f7fa3b0b8345ee046ba36a12d9e530
RUN uv pip install --system --upgrade --force-reinstall --prerelease=allow vllm --extra-index-url https://wheels.vllm.ai/e196028dbaa41b8e3d1c27b7dbd026403af76c27
RUN uv pip install --system --no-cache-dir ".[dev]"
RUN uv pip install --system --upgrade \