Taichang Zhou 7a68bd37b7 Refactor cu_seqlens handling in Qwen3Omni_VisionTransformer to align with upstream vLLM changes. The sequence length helper now retains cu_seqlens in numpy format and converts it for attention execution, improving compatibility and performance.

2026-03-18 23:23:41 +08:00

..

layers

Dev/rebase v0.15.0 (#1159)

2026-02-03 12:24:45 +08:00

model_loader

[Fix]Ensure HuggingFace downloads complete before initialization. (#1213)

2026-02-10 16:05:12 +08:00

models

Refactor cu_seqlens handling in Qwen3Omni_VisionTransformer to align with upstream vLLM changes. The sequence length helper now retains cu_seqlens in numpy format and converts it for attention execution, improving compatibility and performance.

2026-03-18 23:23:41 +08:00

stage_configs

Add Fish Speech S2 Pro support with online serving and voice cloning (#1798)

2026-03-12 16:20:52 -04:00

stage_input_processors

[Bug][Qwen3TTS][Streaming] remove dynamic initial chunk and only compute on initial request (#1930)

2026-03-16 15:23:09 -04:00

__init__.py

update ar_dit_test

2025-09-28 21:58:48 +08:00

custom_process_mixin.py

[Core] customize process between stages, remove the need of "request_ids" in model forward, rebase to 0.12.0 (#458)

2025-12-24 13:26:33 +00:00