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 14, 05:30 UTC
main
[fix] Update torch version in cpu-build.txt for AArch64/ppc64le and Darwin (#25579)
Fadi Arafeh committed about 1 year ago
Revert "[Performance] Move apply_w8a8_block_fp8_linear to an op class… (#25607)
Tyler Michael Smith committed about 1 year ago
typo: remove duplicate `is` (#25641)
Nicole LiHui 🥜 committed about 1 year ago
Map CwmForCausalLM to llama and LlamaForCausalLM (#25611)
Jacob Kahn committed about 1 year ago
[Misc] Simplify PoolerOutput and move to `v1/outputs` (#25629)
Cyrus Leung committed about 1 year ago
[Bugfix] fix apply_temperature to avoid nan in probs (#24734)
courage17340 committed about 1 year ago
optimize: eliminate duplicate split_enc_dec_inputs calls (#25573)
Nicole LiHui 🥜 committed about 1 year ago
[Model] Add LongCat-Flash (#23991)
XuruiYang committed about 1 year ago
Enable Fbgemm NVFP4 on Dense models (#25609)
Saman A. Pour committed about 1 year ago
[Bug] Dynamo Unsupported due to `BasevLLMParameter.torch_function` calling disabled super() (#25613)
Wentao Ye committed about 1 year ago
[Kernel] Support DCP for Triton backend (#25132)
Wei Wei committed about 1 year ago
[Model] Improve DotsOCRForCausalLM (#25466)
Jee Jee Li committed about 1 year ago
[MISC] replace c10::optional with std::optional (#25602)
Shiyan Deng committed about 1 year ago
Improve `--help` for enhanced user experience (#24903)
Harry Mellor committed about 1 year ago
[Refactor] Use DeepGEMM Col Major TMA Aligned Tensor (#25517)
Wentao Ye committed about 1 year ago
feat: BF16 FlashInfer Fused Cutlass MOE for Hopper and Blackwell Expert Parallel (#25503)
Duncan Moss committed about 1 year ago
[Logging] Improve log for when DeepEP HT disables CUDA Graphs (#25531)
Tyler Michael Smith committed about 1 year ago
[V0 Deprecation] Remove unused classes in attention (#25541)
Woosuk Kwon committed about 1 year ago
[Docs] Enable `fail_on_warning` for the docs build in CI (#25580)
Harry Mellor committed about 1 year ago
Suppress benign cuBLAS warning when capturing cudagraphs with DBO (#25596)
Sage Moore committed about 1 year ago