[Rust][Benchmark] Support HF ShareGPT datasets in multi-turn mode (#51104)

JinYan Su committed about 1 month ago

[MRV2] Run pooling post processing on non-final PP ranks (#56666)

Taneem Ibrahim committed about 1 month ago

[Model Runner V2] Acceptance estimation for adaptive verification (#52228)

Giancarlo Delfin committed about 1 month ago

[Bugfix] Redact credentials from benchmark logs (#56662)

Taneem Ibrahim committed about 1 month ago

[ROCm][DSV4.1][Perf] Fold the mHC post step into the delayed pre projection (#56513)

Fangzhou Ai committed about 1 month ago

[CI] Check target branch freshness before starting CI (#56169)

Andreas Karatzas committed about 1 month ago

[Bugfix][Rust Frontend] Restore prost test dependency (#56866)

Kevin H. Luu committed about 1 month ago

[Doc][Metrics] Fix spec-decode PromQL examples to use the exposed names (#55016)

Michele Campi committed about 1 month ago

[ROCm][DSV4.1][Perf] Stride the DSA decode candidate mask over the live context (#56628)

Fangzhou Ai committed about 1 month ago

[Bugfix][DSA] Write nvfp4_ds_mla from the fused norm+rope kernel (#55538)

sychen52 committed about 1 month ago

[Quantization] Select linear backends per quantization (#51204)

Netanel Haber committed about 1 month ago

[Model] Voxtral Realtime: add support for `CUDAGraphMode.FULL_DECODE_ONLY` (#51167)

Nicolò Lucchesi committed about 1 month ago

[Profiler] Add Proton CUDA graph attribution for MRV2 (#51084)

鐘天楽 committed about 1 month ago

[Perf][DSv4.1] Fold the mHC post block into the delayed pre projection (#56633)

Yongye Zhu committed about 1 month ago

[CI] Initialize ubatch runner in cudagraph unit test (#56808)

Kevin H. Luu committed about 1 month ago

[Bugfix] Make KV cache and MFU log lines backend-neutral (#55650)

Zhiyao(Ian) Yu committed about 1 month ago

[Perf][Kernel][Quantization] Fuse ReLU2 with static FP8 activation quantization (#53793)

Samuel Nordmann committed about 1 month ago

[V1][Metrics] Support Sliding Window Attention (SWA) and hybrid layers in MFU/MBU estimation (#55624)

Thillai Chithambaram committed about 1 month ago

[watermarking hardening] Add e2e test and basic GSM8K quality tests (#56309)

Tomas Ruiz committed about 1 month ago

[Rust Frontend] Forward per-request watermarking controls (#56338)

Nicolò Lucchesi committed about 1 month ago