onnxruntime
QMoE CUDA EP — FP4/FP8/WFP4AFP8 Quantized Mixture-of-Experts + MoE GEMM Refactor
#28467
Merged
Overview
Commits
20
Changes
View On GitHub

QMoE CUDA EP — FP4/FP8/WFP4AFP8 Quantized Mixture-of-Experts + MoE GEMM Refactor #28467

tianleiwu merged 20 commits into main from tlwu/20260511/qmoe_cuda
tianleiwu
tianleiwu QMoE CUDA FP4/FP8/WFP4AFP8 + MoE Refactor
17bd084e
tianleiwu tianleiwu requested a review from copilot-pull-request-reviewer copilot-pull-request-reviewer 71 days ago
copilot-pull-request-reviewer
copilot-pull-request-reviewer commented on 2026-05-12
tianleiwu refine
3b1c4150
tianleiwu skip test if not enabled in build
3256b226
tianleiwu fix build
2fce3c41
tianleiwu update op doc
bf919f28
tianleiwu Merge remote-tracking branch 'origin/main' into tlwu/20260511/qmoe_cuda
6abe932e
tianleiwu tianleiwu marked this pull request as ready for review 67 days ago
tianleiwu fix build
b3143e05
tianleiwu tianleiwu force pushed from bbc0af28 to b3143e05 65 days ago
tianleiwu refine build
6b762cae
tianleiwu Do not link cuda in pybind for Windows
6b8dfc3c
tianleiwu share source filters between cuda and cuda plugin
34a988f8
tianleiwu remove commented code; add license header.
8d5aef5e
tianleiwu remove fc3
8f9f41b7
tianleiwu tianleiwu marked this pull request as draft 64 days ago
tianleiwu remove fc3_global_scale; use float8e8m0; wfp4afp8 blackwell support
d77e10c3
tianleiwu K=128 tile support, epilogue fusion, expanded tile configs for SM90 W…
dc609902
tianleiwu update op docs
16ca7670
tianleiwu tianleiwu force pushed from 6c72e8c1 to 16ca7670 64 days ago
github-advanced-security
github-advanced-security commented on 2026-05-18
tianleiwu allow test cuda plugin
b945a10f
github-advanced-security
github-advanced-security commented on 2026-05-18
tianleiwu lintrunner
2580eaa7
tianleiwu clean up
aadd7b0b
tianleiwu tianleiwu marked this pull request as ready for review 63 days ago
tianleiwu tianleiwu requested a review from apsonawane apsonawane 63 days ago
tianleiwu tianleiwu requested a review from kunal-vaishnavi kunal-vaishnavi 63 days ago
tianleiwu remove unused code
fb923805
tianleiwu change testing cuda plugin default to be False
fc18a8e5
kunal-vaishnavi
kunal-vaishnavi commented on 2026-05-20
kunal-vaishnavi
kunal-vaishnavi commented on 2026-05-20
kunal-vaishnavi
kunal-vaishnavi commented on 2026-05-20
kunal-vaishnavi
kunal-vaishnavi commented on 2026-05-20
kunal-vaishnavi
kunal-vaishnavi approved these changes on 2026-05-20
apsonawane
apsonawane commented on 2026-05-20
apsonawane
apsonawane commented on 2026-05-20
apsonawane
apsonawane commented on 2026-05-20
apsonawane
apsonawane commented on 2026-05-20
tianleiwu tianleiwu enabled auto-merge (squash) 62 days ago
tianleiwu tianleiwu merged 548ab6e5 into main 62 days ago
tianleiwu tianleiwu deleted the tlwu/20260511/qmoe_cuda branch 62 days ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
None yet
Milestone
No milestone
Terms Of Use Privacy Policy Refund Policy Impressum