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 7, 21:30 UTC
main
[ModelBash][DSV3] Add TRTLLM DSV3 Router GEMM kernel (6% B1 Speedup) (#34302)
Robert Shaw committed 9 months ago
[XPU] allow TORCH_SDPA/TRITON_ATTN as XPU vit Backend (#35010)
Yan Ma committed 9 months ago
[ROCm][CI] Fix spec decode profile assertion and logprob test determinism (#35043)
Andreas Karatzas committed 9 months ago
[CLEANING] Remove unused disable_by_batch_size from SpeculativeConfig (#35060)
Vincent Gimenes committed 9 months ago
[Refactor] Remove dead private func `_fp8_perm` and `_extract_mask_for_item` (#35068)
Wentao Ye committed 9 months ago
Fix pipeline parallel with embed scaling in the Transformers modelling backend (#35094)
Harry Mellor committed 9 months ago
[Feature] Lazy import for the "mistral" tokenizer module. (#34651)
Neil Schemenauer committed 9 months ago
fix: Apply embedding_multiplier to inputs_embeds (#34813)
Gabe Goodhart committed 9 months ago
[BugFix]: Fix local mypy issues (#34739)
Martin Hickey committed 9 months ago
[Bugfix] Fix kernel benchmark (#33752)
Jee Jee Li committed 9 months ago
[Refactor] Simplify dummy data generation (#35025)
Cyrus Leung committed 9 months ago
[Model Runner V2] Remove propose_draft method (#35070)
Woosuk Kwon committed 9 months ago
[Model Runner V2][Minor] Remove redundant `do_spec_decode` field (#35039)
Nick Hill committed 9 months ago
[Spec Decode] Reduce TP communication for speculative decoding draft token generation (#34049)
qizixi committed 9 months ago
[Model] Add NVFP4 quantization support for Step3.5-Flash (#34478)
tacos8me committed 9 months ago
[Bug] Refactor max_num_batched_tokens to account for drafting (#34898)
Benjamin Chislett committed 9 months ago
qizixi committed 9 months ago
[ROCm][CI] Fix realtime test timeouts caused by aiter JIT compilation delays (#35052)
Andreas Karatzas committed 9 months ago
[ROCm][CI] Fix flaky embedding chat test by using tolerance-based comparison (#35050)
Andreas Karatzas committed 9 months ago
[Bugfix] Fix Qwen3/Qwen3.5 Reasoning Parser (#34779)
Roger Wang committed 9 months ago