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 4, 21:15 UTC
main
[Model Runner V2] Multiple prompt logprobs support (#39937)
Wentao Ye committed 7 months ago
Revert #38730 and #38791 (#40032)
Vadim Gimpelson committed 7 months ago
[UX] Bump version in CG memory profiling log message (#40465)
Matthew Bonanni committed 7 months ago
roikoren755 committed 7 months ago
Shanshan Shen committed 7 months ago
[XPU][CI] Add misc, engine and lora cases on Intel GPU in CI (#39887)
xiangdong committed 7 months ago
[Bugfix] LoRA: extend expert base_layer loading to Qwen3.5 and Step3.x (#37114)
ℍ𝕠𝕝𝕝𝕠𝕨 𝕄𝕒𝕟 committed 7 months ago
[Doc] Add Qwen3 AWQ models to documentation (#40034)
Yusuf Mohammad committed 7 months ago
[Bugfix] Fix dataset name and path argument validation bug in vllm bench serve (#40288)
Talor Abramovich committed 7 months ago
Add Granite 4.1 Vision as built-in multimodal model (#40282)
artem-spector committed 7 months ago
[Frontend] Remove frontend pooling multi task support. (#37861)
wang.yuqi committed 7 months ago
feat(multimodal): support externally processed mm_kwargs with cache injection (#39502)
Kris Hung committed 7 months ago
fix: clamp NaN/Inf in topk_softmax to prevent duplicate expert IDs (#39391)
Jhao-Ting Chen committed 7 months ago
wang.yuqi committed 7 months ago
[Bugfix] Normalize malformed dict prompts that carry token IDs in `prompt` (#40339)
Zeyu Zhang committed 7 months ago
[Feat] dflash support for ROCm (#39703)
hangy-amd committed 7 months ago
[MoE] Triton MoE Perf regression - restore low latency path (#39016)
milesial committed 7 months ago
[MM][Misc] Support image+video mixed inputs (per prompt) for VLM examples (#40335)
Shanshan Shen committed 7 months ago
[Fix] Add missing space in IP fallback warning (#40359)
SeongJun Lee committed 7 months ago
[Deprecation] Deprecate cprofile and cprofile_context (#39100)
Wentao Ye committed 7 months ago