update SP+EP#209

Open

sfc-gh-mhidayetoglu wants to merge 33 commits into

merth/moe_epfrom

Commits on Sep 16, 2025

Reorganize and refactor Suffix Decoding (#182)
sfc-gh-aqiao
authored

Commits on Sep 18, 2025

Commits on Sep 23, 2025

Fix hybrid mode(spec decoding + suffix) crash on structured_output (#169)
sfc-gh-yewang
authored
Make Arctic Inference plugin opt-in instead of opt-out (#188)
sfc-gh-aqiao
authored

Commits on Sep 25, 2025

Simplify min_score selection logic, correct type hint for `propose_suffix_draft_token_ids` (#195)
CptTZ
authored
Add op_builder for jitting the kernels (#193)

sfc-gh-reyazda
and
sfc-gh-aqiao
authored

Commits on Sep 26, 2025

Commits on Oct 6, 2025

Explicitly initialize CUDA buffers for next tokens (#201)
sfc-gh-yewang
authored

Commits on Oct 13, 2025

Commits on Oct 17, 2025

Commits on Oct 20, 2025

More suffix decoding optimizations (#212)
sfc-gh-aqiao
authored

Commits on Oct 21, 2025

Commits on Nov 6, 2025

Commits on Nov 19, 2025

Communication Fusing (#224)
sfc-gh-mhidayetoglu
authored

Commits on Dec 5, 2025

patch for running traces with timestamps (#228)
sfc-gh-mhidayetoglu
authored

Commits on Dec 30, 2025

rebase to vllm 0.11.0 (#216)

sfc-gh-yewang
and
sfc-gh-mhidayetoglu
authored

Commits on Jan 24, 2026

Commits on Jan 26, 2026

Reproducibility extension (#239)
sfc-gh-mhidayetoglu
authored

Commits on Feb 8, 2026

Enhance ArcticProposer and ArcticSpeculator to incorporate scheduler … (#247)
sfc-gh-yewang
authored

Commits on Feb 24, 2026

Rebase to vLLM 0.14.1 (#242)

authored

Commits on Mar 28, 2026

Fix shift-parallel CUDA graph capture and dispatch (#258)

xylian86
and
Mock User
authored