onnxruntime
[CUDA] Vectorize the NVFP4 weight dequantization for prefill
#32128

Merged

[CUDA] Vectorize the NVFP4 weight dequantization for prefill #32128

tianleiwu merged 2 commits into microsoft:main from tianleiwu:tlwu/fp4_prefill_dequant_vec

[CUDA] Vectorize the NVFP4 weight dequantization for prefill

ca8d3947

tianleiwu requested a review from

copilot-pull-request-reviewer 33 days ago

copilot-pull-request-reviewer commented on 2026-08-17

fix(cuda): address fp4 dequant review feedback

54fd1fc5

tianleiwu requested a review from

baijumeswani 31 days ago

tianleiwu requested a review from

hariharans29 31 days ago

tianleiwu requested a review from

titaiwangms 31 days ago

baijumeswani approved these changes on 2026-08-18

hariharans29 approved these changes on 2026-08-18

tianleiwu merged 2d211f8c into main 31 days ago

Reviewers

hariharans29

baijumeswani

titaiwangms

Assignees

No one assigned

Labels

None yet

Milestone

No milestone