[triton] Support head_dim not 2^n in triton extend and decode attention #134
pr-test.yml
on: pull_request
unit-test-frontend
1m 46s
unit-test-backend-part-0
11m 12s
unit-test-backend-part-1
9m 10s
performance-test-1-gpu
13m 15s
performance-test-2-gpu
9m 11s
accuracy-test-1-gpu
4m 6s
accuracy-test-2-gpu
5m 36s
finish
0s
Annotations
7 warnings