onnxruntime
[CUDA] PagedAttention: use exact max_query_len on FA path
#28409

Open

[CUDA] PagedAttention: use exact max_query_len on FA path #28409

elwhyjay wants to merge 2 commits into microsoft:main from elwhyjay:feature/paged-attention-rotary-max-seqlen-fix

elwhyjay force pushed from b80299a1 to 11596b3d 2 days ago

elwhyjay marked this pull request as draft 2 days ago

elwhyjay force pushed from 9e9ba43f to 8278551d 2 days ago

elwhyjay changed the title ~~[CUDA] PagedAttention: use token_count for FA rotary grid~~ [CUDA] PagedAttention: use exact max_query_len on FA path 2 days ago

elwhyjay closed this 2 days ago

elwhyjay reopened this 2 days ago

elwhyjay marked this pull request as ready for review 2 days ago

[CUDA] PagedAttention: use exact max_query_len on FA path

90c5702d

elwhyjay force pushed from 8278551d to 90c5702d 2 days ago

tianleiwu requested a review from

copilot-pull-request-reviewer 2 days ago

copilot-pull-request-reviewer commented on 2026-05-08

Potential fix for pull request finding

37adb160

Reviewers

copilot-pull-request-reviewer

Assignees

No one assigned

Labels

None yet

Milestone

No milestone

onnxruntime [CUDA] PagedAttention: use exact max_query_len on FA path #28409 Open

[CUDA] PagedAttention: use exact max_query_len on FA path #28409

onnxruntime
[CUDA] PagedAttention: use exact max_query_len on FA path
#28409

Open