A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 4, 07:30 UTC
main
[Bugfix] KimiK2ReasoningParser: guard against buffered end-token in streaming (#41068)
Keyi Li committed 6 months ago
Limit gpu utils and lower max BS on test_transcription_api_correctness.py (#41649)
Ekagra Ranjan committed 6 months ago
Wentao Ye committed 6 months ago
Test nemotron nano-v2 and nemotron nano-v3 separately, disable super-omni redundant tests (#41616)
Netanel Haber committed 6 months ago
Revert "[Doc] Fix RTD build: pytorch.org/docs/stable/objects.inv returns 404" (#41618)
Stefano Castagnetta committed 6 months ago
[Bugfix] Apply ruff-format to hyperclovax.py (#41620)
Stefano Castagnetta committed 6 months ago
[CI] Stabilize cpu offload compressed tensors test (#41102)
Andreas Karatzas committed 6 months ago
[ROCm][CI] Align spec decode logprob test prefill settings (#41335)
Andreas Karatzas committed 6 months ago
[Transformers v5] Vendor HCXVisionConfig for compatibility (#38447)
Fang Han committed 6 months ago
[CI] Clean up remote servers on pytest parent exit (#41570)
Andreas Karatzas committed 6 months ago
[Bench] Forward --seed to CustomDataset and CustomMMDataset shuffle (#40788)
Akim Tsvigun committed 6 months ago
[Bugfix] Fix degenerate KV cache stride causing TMA cudaErrorIllegalInstruction (#40737)
David Oy committed 6 months ago
[Bugfix] Fix FP8 Bias Loading (#41424)
Alex Brooks committed 6 months ago
[Bugfix][Ray] Fix RayExecutorV2 actor name collision with DP > 1 (#40398)
tomeras91 committed 6 months ago
Disable flashinfer autotune temporarily due to correctness issues (#41524)
Wei Zhao committed 6 months ago
[Doc] Add Qwen3-30B-A3B-Thinking-2507-FP8 to batch invariance verified models (#41513)
Taneem Ibrahim committed 6 months ago
[MRV2] Add shutdown() method (#41297)
Woosuk Kwon committed 6 months ago
[Quantization] add humming mxfp4 moe backend (#41083)
Jinzhen Lin committed 6 months ago
[DSv4] Tune default value of `VLLM_MULTI_STREAM_GEMM_TOKEN_THRESHOLD` (#41526)
Roger Wang committed 6 months ago
[DSV4] Guard megamoe flag with Pure TP (#41522)
Yongye Zhu committed 6 months ago