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
[BUG] Qwen3-next MTP. Fix attn metadata build bug (#26564)
Vadim Gimpelson committed about 1 year ago
Fix some typing issues found by `mypy==1.18.2` (#26596)
Harry Mellor committed about 1 year ago
Update CUDA architecture list in build pipeline for 12.9.1 wheels (#26592)
Will Eaton committed about 1 year ago
Update `pre-commit` hook versions (#26591)
Harry Mellor committed about 1 year ago
Add Qwen3-Omni moe thinker (#25550)
Xiong Wang committed about 1 year ago
[Transform] [Quantization] Add QuTLASS support to vLLM (#24440)
Roberto L. Castro committed about 1 year ago
[Model] Add FlexOlmo model implementation (#24923)
Shane A committed about 1 year ago
[Model][Qwen3VL] Compute `cu_seqlens` on CPU to remove (#26496)
Lukas Geiger committed about 1 year ago
[Bugfix] fixed top_logprobs: -1 does not appear to work as intended (#26470)
Chauncey committed about 1 year ago
[torch.compile] Make inductor partition rules respect splitting_ops #25691 (#25845)
baonudesifeizhai committed about 1 year ago
[Metrics] Add test for multi-modal cache stats logging (#26588)
Mark McLoughlin committed about 1 year ago
Elvir Crnčević committed about 1 year ago
[Bugfix] Make DP padding optional in coordinate_batch_across_dp (#26375)
Sage Moore committed about 1 year ago
Added test_top_k_per_row to test-pipeline.yaml. (#26569)
Daniel Cámpora committed about 1 year ago
[BugFix][torch.compile] Fix fused_scaled_matmul_reduce_scatter signature for PyTorch 2.8 (#26038)
Jason Li committed about 1 year ago
[CI] fix test_run_batch.py::test_completions - AssertionError (#26578)
Chauncey committed about 1 year ago
[BugFix] Fix noop elimination edge case (#26394)
Andy Lo committed about 1 year ago
[Bugfix] Convert untraceable GroupShape to list for AMD impl (#26535)
Lucas Kabela committed about 1 year ago
[NIXL] Ignore abort on already-finished request (#25067)
Mark McLoughlin committed about 1 year ago
[bugfix][DCP] fix block_size of hash in DCP prefix caching (#26296)
Chen Zhang committed about 1 year ago