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 2, 12:15 UTC
main
[MM][CG] Avoid over-padding Qwen2.5-VL encoder cudagraph window metadata (#42796)
Hua Huang committed 5 months ago
[ROCm] Enable the aiter top-k/top-p sampler by default (#43331)
JohnQinAMD committed 5 months ago
[Attention][AMD] Standardize kv layout to blocks first for AMD (#43660)
Nicolò Lucchesi committed 5 months ago
Allow native KV cache dtype in Triton cache update (#43330)
Mike G committed 5 months ago
[Rust Frontend] Optimize multimodal prompt expansion (#43670)
Chao-Ju Chen committed 5 months ago
Add Cosmos3 Reasoner model (#43356)
MaciejBalaNV committed 5 months ago
[ROCm] Bump ROCm to 7.2.3 (#43136)
Micah Williamson committed 5 months ago
[XPU][MoE] Add WNA16 oracle backend for GPTQ sym-int4 (xpu_fused_moe) (#41426)
Jason Elie Bou Kheir committed 5 months ago
[ROCm][Perf] Support N=5 in wvSplitK skinny GEMM kernels for speculative decoding (#40687)
Matthias Gehre committed 5 months ago
[KV Offload] Rename `SecondaryTierManager.get_finished()` to `get_finished_jobs()` (#43870)
Ronen Schaffer committed 5 months ago
[Bugfix] Exclude Ray DP from #42585's deferred port allocation (#43864)
Vadim Gimpelson committed 5 months ago
[Perf] Add do_not_specialize to Mamba SSD chunk kernels (#43803)
Majid committed 5 months ago
[Bug] Fix `tests/distributed/test_elastic_ep.py - assert False` (#43813)
Wentao Ye committed 5 months ago
[rust] fix: aggregate `is_sleeping` and `reset_prefix_cache` across DP engines (#43429)
Will.hou committed 5 months ago
[Rust Frontend] Reduce Gemma4 tool parser args scan complexity (#43850)
Bugen Zhao committed 5 months ago
[Rust Frontend] Add `hy_v3` tool parser (#43872)
Bugen Zhao committed 5 months ago
[CPU] Migrate cpu_awq into awq_marlin (#43841)
Li, Jiang committed 5 months ago
[Bugfix][ROCm] Resolve MoRI connector hangs at high concurrency (#40344)
Simon Danielsson committed 5 months ago
[Model Refactoring] Remove torch compile dependency in DSv4 (#43746)
Woosuk Kwon committed 5 months ago
Add token-offset based selective offload in OffloadConnector (#39983)
Angelo Ruocco committed 5 months ago