Zhou TaichangandClaude Opus 5 651b03225b [Rebase] Rebase to vllm 0.28.0 (#6606) ...

Signed-off-by: tzhouam <tzhouam@connect.ust.hk>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>

2026-08-27 06:12:55 +00:00

..

__init__.py

[Feature] support multimodal inputs with multiple requests (#76)

2025-11-24 16:02:44 +08:00

audio_length.py

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

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

pipeline.py

[Refactor][2/N Scheduler]Replace _FULL_PAYLOAD_INPUT_STAGES with a resolved stage transport capability. (#6149)

2026-08-25 15:24:59 +08:00

qwen2_5_omni_talker.py

[Perf] Remove dead audio_tower and visual from Qwen2.5-Omni talker stage (#3425)

2026-05-12 23:24:13 +08:00

qwen2_5_omni_thinker.py

[Rebase] Rebase to vllm 0.27.0 (#5976)

2026-08-11 15:23:46 +00:00

qwen2_5_omni_token2wav.py

[Refactor] Unify Snake/SnakeBeta and alias-free activation into common modules (#3886)

2026-05-27 23:41:35 -05:00

qwen2_5_omni.py

[Rebase] Rebase to vllm 0.28.0 (#6606)

2026-08-27 06:12:55 +00:00

qwen2_old.py

[BugFix] Use KV cache scale mapper in custom loaders (#4810)

2026-07-03 04:11:28 +00:00