[Perf] [Bugfix] Fix Triton autotuning in inference for Qwen3.5 (#37338)

Artem Perevedentsev committed 8 months ago

[Bugfix] Store Qwen3Next A_log in fp32 (#37810)

Hojin Yang committed 8 months ago

[Bugfix] JAIS: Only apply ALiBi when position_embedding_type='alibi' (#37820)

r266-tech committed 8 months ago

[ROCm][Refactor] Enable AWQMarlinConfig on ROCm to use choose_mp_linear_kernel (#36505)

Matthias Gehre committed 8 months ago

update doc for online fp8 quantization (#37851)

Yan Ma committed 8 months ago

[Feature] ViT Full CUDA Graph (#35963)

Baorun (Lauren) Mu committed 8 months ago

[CI/Build][LoRA] Update Qwen35 LoRA testing (#37816)

Jee Jee Li committed 8 months ago

[Test] Consolidate tool parser unit tests to tests/tool_parsers (#37834)

Ben Browning committed 8 months ago

always use `embed&token_classify` for bge-m3 (#37632)

Augusto Yao committed 8 months ago

Fix AudioFlamingo3/MusicFlamingo HF parity and RoTE handling (#37643)

Lasha Koroshinadze committed 8 months ago

[MRV2] Enable PP CUDA graph test (#37830)

Woosuk Kwon committed 8 months ago

[Model Runner V2] Enable piecewise CUDA graphs for pipeline parallelism (#35162)

zhanqiuhu committed 8 months ago

[MRV2] Use FP64 for Gumbel noise (#37798)

Woosuk Kwon committed 8 months ago

[Bigfix]fix lora test by pass padded size back to the layer (#37811)

Yongye Zhu committed 8 months ago

[MRV2] Skip hidden states allocation for PW CUDA graphs (#37818)

Woosuk Kwon committed 8 months ago

Enable `NemotronHPuzzle` + `NemotronHMTP` (#37803)

Netanel Haber committed 8 months ago

[Test] Only Run MLA model when user explicitly set for batch invariance (#37719)

Wentao Ye committed 8 months ago

[Bug] Fix fp8 deepgemm batch invariant (#37718)

Wentao Ye committed 8 months ago

[Model Runner V2] Support multi-modal embeddings for spec decode model (#36097)

Giancarlo Delfin committed 8 months ago

[ROCm][CI] Stabilize ROCm speech-to-text translation test with lower min acc threshold (#37723)

Andreas Karatzas committed 8 months ago