SIGN IN SIGN UP

[Operator Mechanism ] Add compatible logsumexp cuda kernel (#79732)

* align logsumexp cuda kernel with torch2.12 when flag is set

* Add tests in test_logsumexp.py to cover logsumexp compatible padth
O
omoYang committed
4f2c7252c138680140d9f61e3fdff66fd5c2ac17
Parent: 3ae8329
Committed by GitHub <noreply@github.com> on 9/10/2026, 3:26:08 AM