[MTP] Refactor mtp predictor to avoid d2h operation (#27643)

Mengqing Cao committed about 1 year ago

Reapply "Install pre-built xformers-0.0.32.post2 built with pt-2.9.0" (#27768)

Huy Do committed about 1 year ago

[Docs] add Shanghai Meetup - 2025/10 (#27545)

Kebe committed about 1 year ago

[Misc] Replace CUDA_VISIBLE_DEVICES in DP with torch.cuda.set_device for device selection on cuda-like devices (#27564)

Ilya Markov committed about 1 year ago

[BugFix] Stopgap - Flashinfer Autotuner + GPT-OSS + DP/TP (#27762)

Varun Sundar Rabindranath committed about 1 year ago

[Bugfix][CPU] Fix MRoPE dispatch on the CPU backend (#27800)

Li, Jiang committed about 1 year ago

[Model][Ouro] Support Ouro Model (#27794)

Fan Yin committed about 1 year ago

[KV offload] Enable CPU KV offload on CUDA alike Platforms (#27770)

Zhewen Li committed about 1 year ago

[Model] Introduce Kimi Linear to vLLM (#27809)

Zhiyuan Li committed about 1 year ago

[CI Failure] Fix test_kv_cache_model_load_and_run (#27717)

Huamin Li committed about 1 year ago

[Frontend][Doc][5/N] Improve all pooling task | Polish encode (pooling) api & Document. (#25524)

wang.yuqi committed about 1 year ago

[Bugfix] Improve GPU validation logging in Ray fallback scenarios (#25775)

Sairam Pillai committed about 1 year ago

[CI] Fix mypy for `vllm/v1/core` and `vllm/v1/engine` (#27108)

Wentao Ye committed about 1 year ago

[CI Failure] fix test_default_mm_loras (#27795)

Huamin Li committed about 1 year ago

[V0 deprecation] Remove VLLM_USE_V1 usage in config module (#27784)

wangxiyuan committed about 1 year ago

[BugFix][VL] Fix FA selection on Qwen2.5-VL (#27790)

Zhewen Li committed about 1 year ago

[CI/Build]Add eval config for Qwen3-235B-A22B-Instruct-2507-FP8 (#27113)

Huamin Li committed about 1 year ago

use stringData in secret yaml to store huggingface token (#25685)

yitingdc committed about 1 year ago

Add more dims for batch invariant shims (#27489)

Bram Wasti committed about 1 year ago

[Benchmark] Cleanup deprecated nightly benchmark and adjust the docstring for performance benchmark (#25786)

Kuntai Du committed about 1 year ago