vllm
[Beam] Enable CuTeDSL NVFP4 MegaMoE
#62073
Open
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Overview
Commits
6
Changes
View On
GitHub
[Beam] Enable CuTeDSL NVFP4 MegaMoE
#62073
WoosukKwon
wants to merge 6 commits into
main
from
woosuk/beam-moe
WoosukKwon
requested a review
from
bnellnm
35 days ago
WoosukKwon
requested a review
from
mgoin
35 days ago
WoosukKwon
requested a review
from
pavanimajety
35 days ago
WoosukKwon
requested a review
from
zyongye
35 days ago
WoosukKwon
requested a review
from
wzhao18
35 days ago
claude
commented on 2026-10-16
mergify
added
nvidia
WoosukKwon
added
beam
WoosukKwon
force pushed
from
cc997c1e
to
b7049100
35 days ago
WoosukKwon
requested a review
from
gau-nernst
35 days ago
WoosukKwon
requested a review
from
vadiklyutiy
35 days ago
WoosukKwon
requested a review
from
DarkLight1337
35 days ago
WoosukKwon
requested a review
from
ywang96
35 days ago
WoosukKwon
requested a review
from
AndreasKaratzas
35 days ago
WoosukKwon
requested a review
from
tlrmchlsmth
35 days ago
WoosukKwon
requested a review
from
yewentao256
35 days ago
WoosukKwon
requested a review
from
khluu
35 days ago
WoosukKwon
requested a review
from
Harry-Chen
35 days ago
WoosukKwon
requested a review
from
njhill
35 days ago
WoosukKwon
requested a review
from
youkaichao
35 days ago
WoosukKwon
requested a review
from
robertgshaw2-redhat
35 days ago
WoosukKwon
requested a review
from
houseroad
35 days ago
WoosukKwon
requested a review
from
ProExpertProg
35 days ago
mergify
added
documentation
mergify
added
ci/build
mergify
added
new-model
mergify
added
qwen
mergify
added
speculative-decoding
mergify
added
dflash
mergify
added
mrv2
WoosukKwon
changed the title
[MoE] Defer MegaMoE finalize under EP and add a decode workspace
[Beam] Enable CuTeDSL NVFP4 MegaMoE
35 days ago
WoosukKwon
force pushed
from
b7049100
to
6fa73f4b
35 days ago
WoosukKwon
force pushed
from
6fa73f4b
to
de488b40
35 days ago
mergify
added
needs-rebase
[Model][Perf] Fuse Beam's TP all-reduce with its sandwich norms and M…
4fb4e753
WoosukKwon
force pushed
from
de488b40
to
86a5bbe7
33 days ago
WoosukKwon
changed the base branch from
main
to
woosuk/beam-fused-allreduce
33 days ago
mergify
removed
needs-rebase
[Model][Perf] Build Beam's fused all-reduce ahead of time under csrc/…
5d683553
[Model][Perf] Note FlashInfer v0.7.2 as the vendored MNNVL source
3ec2cf25
[Model][Perf] Keep Beam's deferred MoE finalize TP-only
9fdc9ed5
mergify
added
needs-rebase
[Build] Gate Beam's csrc kernels under one CMake switch
df583a4b
[Beam] Enable CuTeDSL NVFP4 MegaMoE
81a67076
WoosukKwon
force pushed
from
86a5bbe7
to
81a67076
30 days ago
WoosukKwon
requested a review
from
LucasWilkinson
30 days ago
mergify
removed
needs-rebase
Base automatically changed from
woosuk/beam-fused-allreduce
to
main
30 days ago
mergify
added
needs-rebase
Login to write a write a comment.
Login via GitHub
Reviewers
claude
bnellnm
mgoin
pavanimajety
zyongye
wzhao18
gau-nernst
vadiklyutiy
DarkLight1337
ywang96
AndreasKaratzas
tlrmchlsmth
yewentao256
khluu
Harry-Chen
njhill
youkaichao
robertgshaw2-redhat
houseroad
ProExpertProg
LucasWilkinson
Assignees
No one assigned
Labels
documentation
new-model
speculative-decoding
needs-rebase
ci/build
qwen
nvidia
mrv2
dflash
beam
Milestone
No milestone
Login to write a write a comment.
Login via GitHub