PR Test

[Fix] Adjust default chunked prefill size and cuda graph max bs according to GPU memory capacity #1927

Sign in to view logs

Run time

Learn about OS pricing on GitHub Actions

Job	Run time
accuracy-test-1-gpu	5m 42s
performance-test-1-gpu-part-1	10m 34s
performance-test-1-gpu-part-2	13m 10s
accuracy-test-2-gpu	15m 6s
performance-test-2-gpu	9m 5s
unit-test-backend-part-1	15m 5s
unit-test-backend-part-2	12m 48s
unit-test-backend-part-3	15m 9s
unit-test-backend-part-4	15m 10s
unit-test-frontend	3m 16s
finish	1s
	1h 55m 6s