A high-throughput and memory-efficient inference and serving engine for LLMs
main
[Tools] vLLM Recipes conversion : support different data types variants and strategies (#51878)
Louie Tsai committed 2 months ago
[ROCm][CI] Speed Up ROCm Skinny GEMM Tests (reduced parameterizations, (#51877)
Micah Williamson committed 2 months ago
[Perf][ROCm] Dual-stream decode with hipgraphs (#48223)
Simon Danielsson committed 2 months ago
[ROCm] Enable DeepSeek-V4 on gfx11 (#47017)
PikaPikachu committed 2 months ago
[Bugfix][MRV2] Require all requests to be decoding for uniform-decode dispatch (#51865)
Nick Hill committed 2 months ago
[Attention] Add FlashInfer XQA decode support on SM12x (#49718)
Andrii Skliar committed 2 months ago
[Profiler] Add minimal Triton Proton profiling backend (#48789)
鐘天楽 committed 2 months ago
Lucas Wilkinson committed 2 months ago
[Testing] Fix test_sharded_state_loader (#51736)
Richard Zou committed 2 months ago
[Bugfix][ROCm] Give KV-first attention blocks their own page in hybrid models (#51837)
stefankoncarevic committed 2 months ago
[ROCm][MoE] Fix expert_map vs AITER expert_mask for non-AITER experts under EP (#49758)
Rohan Potdar committed 2 months ago
[Bugfix][MoE] Fix fused block-scale orientation (#50727)
Andreas Karatzas committed 2 months ago
[Refactor] Delete dead code in models (#51838)
Wentao Ye committed 2 months ago
[Bugfix] Support HF-config compat for Inkling (#51850)
Michael Goin committed 2 months ago
[Docs] Fix broken autorefs cross-reference in TurboQuant v2 docstring (#51857)
Harry Mellor committed 2 months ago
[CI][Bugfix][V1] Remove stale FlashAttention metadata arguments (#51854)
Andreas Karatzas committed 2 months ago
[ROCm] Remove stale SDPA and skinny GEMM workarounds (#50907)
Andreas Karatzas committed 2 months ago
[Bugfix][Attention] Forward per-head FP8 descales through FA4 (#51363)
Yi Liu committed 2 months ago
feat: allow shared expert overlapping for FlashInfer one-sided all-to-all (#50569)
Julien Debache committed 2 months ago
Enable `MiniCPMV` for vLLM in CI (#45042)
Harry Mellor committed 2 months ago