onnxruntime
[CUDA][WebGPU] Add PackedSparseAttentionIndexerMerge for MTP index sharing
#33224
Open
Open
[CUDA][WebGPU] Add PackedSparseAttentionIndexerMerge for MTP index sharing #33224
kunal-vaishnavi
wants to merge
3 commits into
main
from
kvaishnavi/index-share
Add stable sparse selection merge for CUDA and WebGPU
22e70ae1
Implement flat IndexShare packed sparse indexer reuse
d93d0a81
kunal-vaishnavi
changed the title
Fix IndexShare plugin builds and WebGPU metadata validation
a1017e9c
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Milestone
No milestone
Login to write a write a comment.
Login via GitHub