Skip to content

[SpecDecode] Reduce TP communication for large-vocab draft models speculative decoding - #39419

Merged
benchislett merged 14 commits into
vllm-project:mainfrom
EanWang211123:local_argmax_for_dflash
Jun 10, 2026
Merged

[SpecDecode] Reduce TP communication for large-vocab draft models speculative decoding#39419
benchislett merged 14 commits into
vllm-project:mainfrom
EanWang211123:local_argmax_for_dflash

Commits

  1. Commits on Apr 9, 2026

  2. Commits on Apr 30, 2026

  3. Commits on May 30, 2026

  4. Commits on Jun 1, 2026

  5. Commits on Jun 9, 2026

  6. Commits on Jun 10, 2026