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, 08:30 UTC
main
[Doc][Neuron] Update documentation for Neuron (#18868)
aws-elaineyz committed over 1 year ago
[Bugfix][TPU] fix moe custom kernel import (#18853)
Chengji Yao committed over 1 year ago
Add ability to use CUDAGraphs with use_inductor=False (#17345)
Richard Zou committed over 1 year ago
Prevent the cross-encoder logic from being applied to classification tasks (#18838)
Maximilien de Bayser committed over 1 year ago
Hongxia Yang committed over 1 year ago
[Core] Enable CUDA graphs for DP + All2All kernels (#18724)
Varun Sundar Rabindranath committed over 1 year ago
Remove checks for `None` for fields which should never be `None` (#17985)
Harry Mellor committed over 1 year ago
[Hardware][TPU][V1] Multi-LoRA Optimisations for the V1 TPU backend (#15655)
Akshat Tripathi committed over 1 year ago
[Chore][Spec Decode] Update check NoneType instead of assigning variables (#18836)
Aaron Pham committed over 1 year ago
[V1][Metrics] Remove metrics that were deprecated in 0.8 (#18837)
Mark McLoughlin committed over 1 year ago
[Misc] fix olmoe model layer can't laod in tp gt 1 (#18828)
rongfu.leng committed over 1 year ago
[Chore] update ty configuration (#18839)
Aaron Pham committed over 1 year ago
[Core] Add Lora Support to Beam Search (#18346)
Alex Brooks committed over 1 year ago
decrement server_load on listen for disconnect (#18784)
daniel-salib committed over 1 year ago
[Frontend] add run batch to CLI (#18804)
Reid committed over 1 year ago
Enable Pydantic mypy checks and convert configs to Pydantic dataclasses (#17599)
Harry Mellor committed over 1 year ago
[Platform][Dist] Make torch distributed process group extendable (#18763)
Mengqing Cao committed over 1 year ago
[BugFix] FA2 MLA Accuracy Issue (#18807)
Lucas Wilkinson committed over 1 year ago
Fix PiecewiseCompileInterpreter (#17338)
Richard Zou committed over 1 year ago
[CI] improve embed testing (#18747)
wang.yuqi committed over 1 year ago