[Bugfix] Prevent stale multiproc RPC deadlines from becoming unbounded waits (#41357)

bugkeep committed 3 months ago

[Frontend] Add diarized_json support for MOSS-Transcribe-Diarize (#48543)

christian committed 3 months ago

[CI Bugfix] Temp disable Humming wNa8 INT8 H100 CI (#50329)

Michael Goin committed 3 months ago

[CI][ROCm] Fix AMD nightly distributed regressions (#50304)

Andreas Karatzas committed 3 months ago

[ROCm][CI] Avoid Ray worker startup env race (#50311)

Andreas Karatzas committed 3 months ago

[Rubin] Enable NVLink all-reduce paths on SM107 (#49647)

zaristei committed 3 months ago

feat(grpc): add KV event source discovery (#50033)

Connor Carpenter committed 3 months ago

Revert "[Misc][Minimax-M3]add default video_processor (#50092)" (#50313)

Nick Hill committed 3 months ago

[Perf] RMSNorm uncontiguous support, 1.2~3.1x kernel performance improvement (#49750)

Wentao Ye committed 3 months ago

[torch.compile] Compile `CustomOp.forward_native` for ReLU^2 to avoid raw torch ops inside opaque custom ops (#50244)

roikoren755 committed 3 months ago

[docs] Add documentation for pynvvideocodec video decoding backend (#49066)

Brandon Pelfrey committed 3 months ago

fix(step3p5-mtp): honor exclude_modules for the MTP head via prefix (#48883)

Chanh Nguyen committed 3 months ago

[KV Connector] Fix NIXL mamba state pairing for multi-slot block tables (#50153)

Nick Hill committed 3 months ago

[ROCm][CI] Fix Kimi K3 KDA on ROCm (#50262)

stefankoncarevic committed 3 months ago

[Spec Decode][Perf] Replicate DSpark Markov head across TP ranks (#49731)

Michael Goin committed 3 months ago

[BugFix] Fix `num_output_placeholders` preemption underflow (#48245)

Nick Hill committed 3 months ago

[ModelRunner V2] Enable sequence pooling for embedding and classification models (#48791)

Taneem Ibrahim committed 3 months ago

[Bugfix][Frontend] Return transcription and translation verbose as float (#49073)

christian committed 3 months ago

[Rust Frontend] Send multimodal tensors in auxiliary frames (#49341)

Reid committed 3 months ago

[CI][Test] Fix pooling truncation test after VLLMError hierarchy change (#50241)

stefankoncarevic committed 3 months ago