CI

Revert "perf(cuda): use scalar tensor for attention neg_large instead… #119

Sign in to view logs

Run time

Learn about OS pricing on GitHub Actions

Job	Run time
Test (Android aarch64)	2m 19s
Test (ubuntu-latest)	1m 39s
Test (macos-latest)	1m 25s
Clippy	1m 41s
Test (linux-flash-attn)	46m 18s
Test (windows-latest)	5m 34s
Test (linux-rocm)	1m 12s
Test (linux-cuda)	6m 29s
Test (linux-vulkan)	1m 16s
Test (macos-metal)	1m 46s
Test (iOS arm64)	2m 38s
	1h 12m 17s