@tduncan / Nvidia-Optimization-Stack-VLLM Mirror · read-only
Fork 0

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 15, 23:30 UTC

main

Consolidate Llama model usage in tests (#13094)

Harry Mellor committed almost 2 years ago

[Misc] Remove redundant statements in scheduler.py (#13229)

Wang Ran (汪然) committed almost 2 years ago

[Bugfix] Offline example of disaggregated prefill (#13214)

XiaobingZhang committed almost 2 years ago

[V1] Consolidate MM cache size to vllm.envs (#13239)

Roger Wang committed almost 2 years ago

[ROCm] Avoid using the default stream on ROCm (#13238)

Gregory Shtrasberg committed almost 2 years ago

Revert "Add label if pre-commit passes" (#13242)

Harry Mellor committed almost 2 years ago

Optimize moe_align_block_size for deepseek_v3 (#12850)

Michael Goin committed almost 2 years ago

Add label if pre-commit passes (#12527)

Harry Mellor committed almost 2 years ago

[Misc] Qwen2.5-VL Optimization (#13155)

燃 committed almost 2 years ago

[VLM] Merged multi-modal processor for Molmo (#12966)

Cyrus Leung committed almost 2 years ago

[Frontend] Pass pre-created socket to uvicorn (#13113)

Russell Bryant committed almost 2 years ago

← Newer Page 891 of 1120 Older →