onnxruntime
Clamp derived sequence lengths and KV-cache index in CUDA GroupQueryAttention
#29240

Merged

Clamp derived sequence lengths and KV-cache index in CUDA GroupQueryAttention #29240

jiafatom merged 5 commits into microsoft:main from jiafatom:jiafatom/gqa-cache-index-bounds

Clamp derived sequence lengths and KV-cache index in CUDA GroupQueryA…

30f11f6b

Fix CUDA GQA cache-append test: size present_kv via initializer total…

cd70bc37

Assert shape (not finiteness) for malformed-seqlens_k CUDA GQA test

e33469eb

Drop unnecessary lambda capture in CUDA GQA test verifier

77287e1d

tianleiwu requested a review from

copilot-pull-request-reviewer 17 days ago

copilot-pull-request-reviewer commented on 2026-06-24

Clarify why the malformed-seqlens_k CUDA GQA test verifies shape only

282d7003

apsonawane approved these changes on 2026-06-24

jiafatom merged eade9ea9 into main 16 days ago

jiafatom deleted the jiafatom/gqa-cache-index-bounds branch 16 days ago

Reviewers

apsonawane

Assignees

No one assigned

Labels

None yet

Milestone

No milestone