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
[Multimodal] Update Tensor schema test to cover arbitrary shape mm inputs (#22867)
Isotr0py committed about 1 year ago
[Misc] Add --save-dir option to benchmark_moe (#23020)
Jee Jee Li committed about 1 year ago
Andrew Sansom committed about 1 year ago
[Build] Env var to disable sccache (#22968)
Lucas Wilkinson committed about 1 year ago
[XPU]avoid circular import during XPU init (#23017)
Kunshang Ji committed about 1 year ago
[misc] nsys profile output kernel classifier and visualizer (#22971)
Grace Ho committed about 1 year ago
[V1] support min_tokens for detokener (#22014)
Calvin Chen committed about 1 year ago
[Frontend] Avoid list copies in `serving_chat.py` (#22947)
Nick Hill committed about 1 year ago
[Bugfix] Fix DeepSeek MTP (#22934)
Benjamin Chislett committed about 1 year ago
[Kernel] Simplify `get_kv_cache_layout` and cache `use_trtllm_attention` env-dependent bit (#22735)
Nicolò Lucchesi committed about 1 year ago
[Misc] Support passing multiple request ids at once to `AsyncLLM.abort()` (#22944)
Nick Hill committed about 1 year ago
[BugFix] Make `run_once` thread-safe (#22978)
Yichen Yan committed about 1 year ago
Support multiple attention groups for KV sharing (#22672)
Yong Hoon Shin committed about 1 year ago
[v1] Move block_hashes from KVCacheManager to Request.block_hashes (#19728)
Or Ozeri committed about 1 year ago
[BugFix] Handle case where async utility call is cancelled (#22996)
Nick Hill committed about 1 year ago
rishitdholakia13 committed about 1 year ago
ci: Add CUDA + arm64 release builds (#21201)
Eli Uriegas committed about 1 year ago
[BugFix] Fix regression caused by mamba state dtype PR (#22998)
Thomas Parnell committed about 1 year ago
[Attention] FA3 Attention Sinks Perf Boost (#22478)
Lucas Wilkinson committed about 1 year ago
minor: zero workspace buffer init for flashinfer trtllm-gen attn (#22603)
eigen committed about 1 year ago