rebase: adapt to vLLM 0.29 callback, embedding and video decoder changes

Four unrelated 0.29 surface changes, each reproduced locally against the
pinned f5c3cc24 build:

- PromptUpdateDetails.is_embed now takes only the full token list. The
  Bagel img2img replacement still declared (tok, seq), which crashed
  stage-0 engine core startup and produced the Distributed and Qwen3-Omni
  "no live replica" failures rather than anything KV related.

- dispatch_unquantized_gemm gained a linear_backend argument and is called
  with it. The CPU-forcing test stubs took no arguments; they now accept
  and ignore whatever upstream passes so a further signature change does
  not break them again.

- VocabParallelEmbedding gained a fused kernel that engages whenever
  tp_size > 1 on CUDA, and it has no CPU implementation. The encoder TP
  fixtures drive CPU tensors under a faked TP group, so they pin the
  portable masked path.

- The pyav video decoder backend was removed; valid options are now
  opencv, torchcodec, pynvvideocodec and deepstream. The OpenAI video
  endpoint hardcoded pyav, which turned every video request into a 400.
  It moves to opencv, the upstream default.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0146Xu5Sgj4vXxLxvLjV5YPj
Signed-off-by: tzhouam <tzhouam@connect.ust.hk>
This commit is contained in:
tzhouamandClaude Opus 5 committed 2026-09-10 00:41:53 +08:00
1 parent c306445ff2
commit 08884c4cfb
10 files changed
+26 -9

No files matched your search

@@ -433,7 +433,8 @@ class OmniBagelMultiModalProcessor(BaseMultiModalProcessor[OmniBagelProcessingIn
embed_mask = [True] * num_vae_total + [False] + [True] * num_vit_total
return PromptUpdateDetails(
full=tokens,
is_embed=lambda _tok, _seq, _m=embed_mask: torch.tensor(_m, dtype=torch.bool),
# vLLM 0.29 calls is_embed with just the full token list.
is_embed=lambda _full, _m=embed_mask: torch.tensor(_m, dtype=torch.bool),
)
replacements.append(