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 6, 10:45 UTC
main
fix(worker): optimize swap_states to copy only active token prefixes (#34733)
Philip Ottesen committed 8 months ago
[V0 Deprecation] Deprecate virtual engine (#37195)
Wentao Ye committed 8 months ago
chunk parakeet into 30s clips to prevent OOMs on long audios (#36671)
Netanel Haber committed 8 months ago
[Bugfix] Fix ROCm crash in qwen3_next multi-stream events (#36795) (#37427)
JartX committed 8 months ago
Fix models which use `layer_type_validation` for Transformers v5 (#37398)
Harry Mellor committed 8 months ago
[Bugfix] Fix incorrect use of merge_size in Qwen3-VL video timestamp calculation (#37439)
Chengyu Fang committed 8 months ago
[bugfix][async scheduling] fix extra cuda context in device 0 with EP/DP (#37449)
youkaichao committed 8 months ago
[Model] Remove unnecessary processor definition for Nemotron Parse (#37456)
Cyrus Leung committed 8 months ago
[Misc] Clean up model registry (#37457)
Cyrus Leung committed 8 months ago
[Perf] Add tuned triton moe config for Qwen3.5 H200, 9.9% E2E throughput improvement (#37340)
Wentao Ye committed 8 months ago
[kv_offload+HMA][2/N]: Support multiple KV groups in GPULoadStoreSpec (#36642)
Or Ozeri committed 8 months ago
Add API docs link if the CLI arg is a config class (#37432)
Harry Mellor committed 8 months ago
Adding deterministic lora benchmarking to vLLM Bench (#36057)
RonaldBXu committed 8 months ago
[Kernel] Add gpt-oss Router GEMM kernel (#37205)
Xin Yang committed 8 months ago
standardize load_weights using AutoWeightsLoader for kimi_linear and minimax_text_01 (#37371)
XLiu-2000 committed 8 months ago
[Perf] Eliminate padding and slicing op for GPT-OSS with Flashinfer MXFP4 MXFP8 MoE (#30647)
elvischenv committed 8 months ago
[Log] Reduce duplicate log (#37313)
Wentao Ye committed 8 months ago
elastic_ep: Fix stateless group port races (#36330)
Itay Alroy committed 8 months ago
[2/3] Refactor InternVL-based processors (#37324)
Cyrus Leung committed 8 months ago
[kv_offload+HMA][6/N]: Split offloading_connector.py (#37405)
Or Ozeri committed 8 months ago