[Misc] Removed undefined cmake variables MOE_PERMUTE_ARCHS (#21262)

Yang Chen committed over 1 year ago

[CI/Build] fix cpu_extension for apple silicon (#21195)

Ignacio Sica committed over 1 year ago

[Misc][Tools] make max-model-len a parameter in auto_tune script (#21321)

Chengji Yao committed over 1 year ago

[Model] Fix a check for None but the return value was empty list in Gemma3 MM vision_embeddings (#21479)

hfan committed over 1 year ago

[Model] Support tensor parallel for timm ViT in Deepseek_vl2 (#21494)

Jason Gu committed over 1 year ago

[Bugfix] fix modelscope snapshot_download serialization (#21536)

Ning Xie committed over 1 year ago

[CI] Update CODEOWNERS for CPU and Intel GPU (#21582)

Li, Jiang committed over 1 year ago

Integrate TensorSchema with shape validation for Phi3VImagePixelInputs (#21232)

Benji Beck committed over 1 year ago

[Docs] Add `requirements/common.txt` to run unit tests (#21572)

Zhou Fang committed over 1 year ago

[TPU][Test] Temporarily suspend this MoE model in test_basic.py. (#21560)

QiliangCui committed over 1 year ago

[DP] Support api-server-count > 0 in hybrid DP LB mode (#21510)

Nick Hill committed over 1 year ago

[Bugfix] DeepGemm utils : Fix hardcoded type-cast (#21517)

Varun Sundar Rabindranath committed over 1 year ago

[Kernel] adding fused_moe configs for upcoming granite4 (#21332)

Burkhard Ringlein committed over 1 year ago

Fix GLM-4 PP Missing Layer When using with PP. (#21531)

Yuxuan Zhang committed over 1 year ago

[Bug] Fix DeepGemm Init Error (#21554)

Wentao Ye committed over 1 year ago

[Docs] Fix `site_url` for RunLLM (#21564)

Harry Mellor committed over 1 year ago

[Frontend] `run-batch` supports V1 (#21541)

Cyrus Leung committed over 1 year ago

[MoE] More balanced expert sharding (#21497)

Woosuk Kwon committed over 1 year ago

[TPU][TEST] HF_HUB_DISABLE_XET=1 the test 3. (#21539)

QiliangCui committed over 1 year ago

update flashinfer to v0.2.9rc1 (#21485)

weiliang committed over 1 year ago