[Misc][Harmony] Move Responses API only harmony utils to responses/harmony.py (#35339)

Flora Feng committed 8 months ago

[BugFix] anthropic/serving_messages: fix tool call arguments streaming (#34887)

Daniele committed 8 months ago

[torch.compile] Sequence Parallelism threshold compile ranges (#28672)

Jason Li committed 8 months ago

[CPU][Feat] Enable KleidiAI INT8_W4A8 for all input dtypes (#34890)

Fadi Arafeh committed 8 months ago

[Model Runner V2] Add coding style guide (#35325)

Woosuk Kwon committed 8 months ago

[Kernel] Refactor FlashInfer allreduce for mnnvl backend (#34109)

Hanjie Qiu committed 8 months ago

openpangu-vl support video input (#34134)

hujiaxin0 committed 8 months ago

[BugFix] Fix fp4 quant kernel on CUDA 12.8 (#35210)

Roberto L. Castro committed 8 months ago

[Bugfix] Fix CUDA compatibility path setting for both datacenter and consumer NVIDIA GPUs (#33992)

Seungmin Kim committed 8 months ago

[Bugfix] Fix AttributeError in SMControlContextManager (#35338)

Lucas Wilkinson committed 8 months ago

[UX] Add `--moe-backend` arg for explicit kernel selection (#33807)

Michael Goin committed 8 months ago

[MoE Refactor] MXFP4 Cutlass Experts to MK (#34542)

Yongye Zhu committed 8 months ago

[UX] Add `--performance-mode {balanced,interactivity,throughput}` (#34936)

Michael Goin committed 8 months ago

[offloader] v2: Hide weight onloading latency via prefetching (#29941)

Ming Yang committed 8 months ago

[ROCm][CI] Amending deletion of AMD mirror (#35322)

Andreas Karatzas committed 8 months ago

[ROCm][CI] Extending attention backend coverage for Eagle spec decode tests (#35265)

Andreas Karatzas committed 8 months ago

fix(mxfp4): Disable monolithic path for TRITON backend with EP (#34270)

Elizabeth Thomas committed 8 months ago

[CI][AMD][BugFix] Add torch.cuda.set_device to test_punica_ops so punica kernels execute on same device as tensor (#34985)

rasmith committed 8 months ago

Revert "[Misc] Enable weights loading tracking for quantized models" (#35309)

Lucas Wilkinson committed 9 months ago

[ROCm][CI] Disable skinny GEMMs in multimodal tests to fix non-deterministic results (#35049)

Andreas Karatzas committed 9 months ago