Jiangyun Zhu 0df8e80574 [diffusion] use fa3 by default when device supports it (#783) ...

Signed-off-by: zjy0516 <riverclouds.zhu@qq.com>

2026-01-22 15:00:29 +08:00

..

diffusion

[Diffusion] Non-Intrusive Sequence Parallelism (SP) Model Support Abstraction for vLLM-Omni Framework (#779)

2026-01-21 14:43:09 +08:00

distributed/omni_connectors

[Security][Feature] Use Msgpack in OmniSerializer instead of pickle (#310)

2025-12-22 13:21:16 +13:00

e2e

[diffusion] use fa3 by default when device supports it (#783)

2026-01-22 15:00:29 +08:00

entrypoints

omni entrypoint support tokenizer arg (#572)

2026-01-08 15:44:14 +08:00

model_executor/models/qwen2_5_omni

[Bugfix]Fix Qwen 2.5 omni hardcoded max_mel_frames to resolve shape mismatch (#543)

2026-01-01 17:12:03 +13:00

worker

Dev/rebase 0.14.0 (#847)

2026-01-21 00:25:25 +08:00

__init__.py

[Fix] Resolve comments & update vLLM-Omni name usages. (#122)

2025-12-01 08:05:24 +08:00

conftest.py

[Test] Add full test for Qwen3-Omni-30B-A3B-Instruct for image and audio single modal (#827)

2026-01-21 13:04:46 +08:00

utils.py

[CI] split tests to avoid timeout (#883)

2026-01-21 15:12:45 +08:00