[Compressed Tensors] Allow configs with non-explicit ignores (#41965)

Kyle Sayers committed 6 months ago

Laguna xs dflash support (#41880)

MeganEFlynn committed 6 months ago

[ROCm] Fix AITER AR+RMSNorm no-residual fusion (#41972)

akii96 committed 6 months ago

[ROCm][DeepSeek] Enable V3.2 TP4 AITER MLA (#41835)

akii96 committed 6 months ago

[Core] Replace routing replay with device cache and async D2H pipeline (#39917)

TomerBN-Nvidia committed 6 months ago

[Refactor] Consolidate required/named tool_choice streaming into DelegatingParser (#41876)

Flora Feng committed 6 months ago

[Misc] Delay EPLB Nixl import until needed (#41805)

Nicolò Lucchesi committed 6 months ago

[Model Runner V2] support qwen35 / mamba hybrid model (#35520)

zhrrr committed 6 months ago

Fix spec decode benchmark metrics (#41916)

noobHappylife committed 6 months ago

tokenizer: Add fastokens support (#41741)

AlonKejzman committed 6 months ago

[ROCm][CI] Avoid duplicate ROCm AITER norm-quant patterns (#41534)

Andreas Karatzas committed 6 months ago

[XPU] Implement out-of-place all-reduce functionality (#41808)

Chaojun Zhang committed 6 months ago

[XPU] Fix lora bugs & enable UTs under tests/lora (#38206)

Chaojun Zhang committed 6 months ago

[Bugfix] Fix GLM4-MoE weight loading for NVFP4 quantized checkpoints (#41755)

s-yanev committed 6 months ago

[MM][Gemma4] Use video profiling hints in encoder budget (#41837)

SeongJun Lee committed 6 months ago

[CPU] Bump up to the latest CPU kernels (#41924)

Li, Jiang committed 6 months ago

[Model] Use AutoWeightsLoader for AXK1 (#41901)

wenyili committed 6 months ago

[Bugfix] Fix FusedMoEWithLoRA has no attribute `runner` (#41889)

Jee Jee Li committed 6 months ago

[CI][Arm] skip e2e model tests if HF_TOKEN is not set (#41919)

Fadi Arafeh committed 6 months ago

Eliminate redundant MoE buffer copies in AITER fused experts (without dependency on AITER changes) (#41713)

Mehdi Ghanimifard committed 6 months ago