vllm-omni/vllm_omni/diffusion/cache at codex-api-server-user-guide - vllm-omni - Gitea: Git with a cup of tea

anodyine/vllm-omni

0 Fork 0

Files

T

History

Wenjia LiandWeiQing Chen d9980c75e8 Perf/minimax h3 dit swiglu rope cache npu (#6410 )

Signed-off-by: Wenjia Li <wjialish@gmail.com>
Co-authored-by: WeiQing Chen <40507679+david6666666@users.noreply.github.com>

2026-08-29 09:14:52 +00:00

..

[Model][Feat] Support Minimax-H3 quality grading requests through dynamic loading/unloading with Cache-DiT (#5853 )

2026-08-07 09:07:23 +00:00

[Bugfix][Diffusion] Apply MagCache residual once per skipped step (#5561 )

2026-07-30 07:09:25 -07:00

[Perf][DreamZero]dreamzero supports CUDA graph, torch.compile, and DiT Caching. (#4213 )

2026-06-19 14:24:37 +08:00

Perf/minimax h3 dit swiglu rope cache npu (#6410 )

2026-08-29 09:14:52 +00:00

__init__.py

[Feat] Support MagCache (#1287 )

2026-05-29 22:26:23 +08:00

base.py

[Perf][DreamZero]dreamzero supports CUDA graph, torch.compile, and DiT Caching. (#4213 )

2026-06-19 14:24:37 +08:00

prompt_embed_cache.py

[feat] Support prompt embedding caching for diffusion model (#2962 )

2026-05-29 15:36:03 +08:00

selector.py

[Refactor][1/N]diffusion/cache: cache-dit (#5226 )

2026-07-24 22:42:45 +08:00