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 15, 21:00 UTC
main
[V1][Metrics] Handle preemptions (#13169)
Mark McLoughlin committed over 1 year ago
[ROCm][V1] Update reshape_and_cache to properly work with CUDA graph padding (#13922)
Sage Moore committed over 1 year ago
[misc] Rename Ray ADAG to Compiled Graph (#13928)
Rui Qiao committed over 1 year ago
Use CUDA 12.4 as default for release and nightly wheels (#12098)
Michael Goin committed over 1 year ago
[ROCm][Quantization][Kernel] Use FP8 FNUZ when OCP flag is 0 or undefined (#13851)
ℍ𝕠𝕝𝕝𝕠𝕨 𝕄𝕒𝕟 committed over 1 year ago
[Kernel] FlashMLA integration (#13747)
Lucas Wilkinson committed over 1 year ago
[BugFix] Make FP8 Linear compatible with torch.compile (#13918)
Woosuk Kwon committed over 1 year ago
[Bugfix] Backend option to disable xgrammar any_whitespace (#12744)
Wallas Henrique committed over 1 year ago
[Misc] fixed qwen_vl_utils parameter error (#13906)
Chauncey committed over 1 year ago
[TPU] use torch2.6 with whl package (#13860)
Chenyaaang committed over 1 year ago
[Bugfix] Update expected token counts for Ultravox tests (#13895)
Cyrus Leung committed over 1 year ago
[Bugfix] Do not crash V0 engine on input errors (#13101)
Joe Runde committed over 1 year ago
[Misc]: Add support for goodput on guided benchmarking + TPOT calculation refactor (#13736)
Brayden Zhong committed over 1 year ago
[Bugfix] Handle None parameters in Mistral function calls. (#13786)
Florian Greinacher committed over 1 year ago
Add comments on accessing `kv_cache` and `attn_metadata` (#13887)
Harry Mellor committed over 1 year ago
[Bugfix] Add test example for Ultravox v0.5 (#13890)
Cyrus Leung committed over 1 year ago
[Misc] Fix input processing for Ultravox (#13871)
Roger Wang committed over 1 year ago
[Misc] Improve LoRA spelling (#13831)
Jee Jee Li committed over 1 year ago
[v0][Core] Use xgrammar shared context to avoid copy overhead for offline engine (#13837)
Seth Kimmel committed over 1 year ago
[ROCm] Disable chunked prefill/prefix caching when running MLA on non-cuda platforms (#13844)
Sage Moore committed over 1 year ago