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 9, 01:15 UTC
main
fix memory for online fp8 quantization with streaming weight load (#31914)
Vasiliy Kuznetsov committed 9 months ago
[UX] Format attention backend log line (#33570)
Matthew Bonanni committed 9 months ago
yugong333 committed 9 months ago
Update huggingface-hub again (#33567)
Harry Mellor committed 9 months ago
Remove incorrect tokenizer info test (#33565)
Harry Mellor committed 9 months ago
[Model] Use mm_position to compute mrope positions for GLM-4.xV (#33039)
Yang Liu committed 9 months ago
[CI] Add DeepSeek V3.2 nightly eval (#33566)
Matthew Bonanni committed 9 months ago
[Refactor] Move profiling methods to MM budget (#33559)
Cyrus Leung committed 9 months ago
[Feature][Core] Support Fabric detection to adapt the MNNVL protocol for the GB series (#33540)
Kebe committed 9 months ago
move spec decode slow test to test_areas.yaml (#33365)
shanjiaz committed 9 months ago
[Bugfix] Enable Kimi k25 processor test (#33562)
Isotr0py committed 9 months ago
[MoE] Enable Shared/Routed Overlap For Latent MoE (Nemotron-H) (#32790)
danielafrimi committed 9 months ago
fix[ROCm]: Remove unconditional aiter import (#32902)
Rabi Mishra committed 9 months ago
[Model] Use explicit types in `get_generation_prompt` (#33551)
Cyrus Leung committed 9 months ago
Update get_expert_mapping to include self parameter (#33525)
Borushiki committed 9 months ago
Fix accessing hidden_act from model config (#32686)
Grzegorz K. Karch committed 9 months ago
Nicolò Lucchesi committed 9 months ago
[Chore] Remove redundant input parsing methods (#33542)
Cyrus Leung committed 9 months ago
[Misc] support arbitrary MM datasets in spec dec bench (#33486)
Komal Kumar Teru committed 9 months ago
[CPU][IBM Z][Dockerfile] Fix IBM Z builds (#33243)
R3hankhan committed 9 months ago