Fix/resupport nongated fused moe triton #36412
vllm-bot
merged
24 commits into
vllm-project:main
from
shaunkotek:fix/resupport-nongated-fused-moe-triton
gemini-code-assist
commented
on 2026-03-08
mgoin
approved these changes
on 2026-03-08
resupport non gated fused moe in triton
c1126dba
[Model] Nano Nemotron VL - fast media preprocessing (#35657)
bb679b1a
[Frontend] Add GPU-less render serving path (`vllm launch render`) (#…
570bf720
Add support for ModelOpt MXFP8 MoE models (#35986)
730745c2
[cudagraph] fix cudagraph warning in deepseekv32 (#28044)
5a47a952
[XPU][Doc] update xpu document about triton dependency/conflict issue…
d6ec74d7
Allow `markdownlint` to run locally (#36398)
ad818e78
[Dependency] Remove default ray dependency (#36170)
6a392273
[Bugfix] Fix CPU OMP autobind assertion to use local_world_size (#35815)
6cefd2bb
[Examples][1/n] Resettle basic examples. (#35579)
6ffbf5c6
fix: Use iterator as not to store all the file loads in memory at onc…
fa16442a
Increase Flexibility for OOV Multimodal Token Handling (#34858)
1bca5351
[Misc] Move processors to `transformers_utils` (#35953)
c7ebbe70
feat(attention): extract KV-cache update from FlexAttention backend (…
4fdaf0ad
[Bugfix] Skip out-of-stage layers in get_layers_from_vllm_config for …
18c8fb96
[Frontend][2/n] Improve pooling entrypoints | embed. (#36110)
cda0a73c
[Bugfix] Avoid to replace non-tensor members in cpu model runner (#36…
b3b43160
[Frontend] Add Support for MM Encoder/Decoder Beam Search (Online Tra…
97c8f800
[XPU] Add test script of PD disaggregation (#36434)
a6291dfc
[Kernel] Add fused_sigmoid_gating_delta_rule_update kernel for Qwen3 …
2bfb5759
[Deprecation][1/2] Remove items deprecated in v0.18 (#36470)
aae9f536
[ci] Bound openai dependency to 2.24.0 (#36471)
9bbc5e63
[MM Encoder] Default to use TORCH_SDPA backend for ViT on Volta/Turin…
e4b4a459
Merge branch 'main' into fix/resupport-nongated-fused-moe-triton
3614f5ad
Reviewers
Assignees
No one assigned
Labels
documentation
performance
rocm
frontend
speculative-decoding
ready
ci/build
v1
multi-modality
qwen
cpu
kv-connector
nvidia
Milestone
No milestone
Login to write a write a comment.
Login via GitHub