[Bugfix] Fix GPT-OSS AR+NORM fusion (#28841)

elvischenv committed 12 months ago

[ROCm][CI] Fix test_cudagraph_mode failure in AMD CI (#29367)

Micah Williamson committed 12 months ago

EAGLE Support DP>1 (#26086)

Rémi Delacourt committed 12 months ago

[Perf] use cpu all reduce to avoid sync when async_scheduling & dp > 1 (#29311)

zhrrr committed 12 months ago

Allow oot custom compiler extension via CompilerInterface (#28623)

Icey committed 12 months ago

[CI/Build] Pin torchgeo dependency for AMD (#29353)

Ryan Rock committed 12 months ago

[fix][cpu] Use a SwigluOAI impl which supports interleaved gate-up wei (#29273)

Fadi Arafeh committed 12 months ago

[Bugfix] [ROCm] [UX]: revert Flex attention backend (#29371)

vllmellm committed 12 months ago

[BugFix] Use unique ids for different transcription prompts (#29372)

Nick Hill committed 12 months ago

Display warning only when ROCm version is less than Pytorch required version (#29200)

Inoki committed 12 months ago

Fix PoolingParams.skip_reading_prefix_cache type (#29364)

kflu committed 12 months ago

Add TP CLI argument to multimodal inference examples (#29301)

Fanli Lin committed 12 months ago

Scheduled removal of `guided_*` config fields (#29326)

Harry Mellor committed 12 months ago

[CI/Test Fix] Fix CP tests on Blackwell (#29338)

Lucas Wilkinson committed 12 months ago

[Attention] add `_cudagraph_support` for linear attention (#28934)

Jiangyun Zhu committed 12 months ago

[Model] Add HunyuanOCR support (#29327)

Isotr0py committed 12 months ago

[Bugfix] Make deprecated `--task embedding` consistent with `--runner… (#29312)

Maryam Tahhan committed 12 months ago

[Metrics] Scheduled removal of deprecated metrics (#29330)

Mark McLoughlin committed 12 months ago

[Docker] Optimize Dockerfile: consolidate apt-get and reduce image size by ~200MB (#29060)

汪志鹏 committed 12 months ago

[CI][ROCm] Install arctic-inference on ROCm tests (#29344)

Divakar Verma committed 12 months ago