[Feature] Implement gradient checkpointing (#1319) #543
Job | Run time |
---|---|
4m 45s | |
5m 11s | |
4m 49s | |
4m 28s | |
5m 28s | |
4m 5s | |
4m 17s | |
4m 4s | |
4m 24s | |
5m 33s | |
6m 16s | |
7m 51s | |
20m 41s | |
12m 46s | |
6m 50s | |
18m 2s | |
10m 25s | |
17m 34s | |
21m 8s | |
11m 28s | |
3h 0m 5s |
Job | Run time |
---|---|
4m 45s | |
5m 11s | |
4m 49s | |
4m 28s | |
5m 28s | |
4m 5s | |
4m 17s | |
4m 4s | |
4m 24s | |
5m 33s | |
6m 16s | |
7m 51s | |
20m 41s | |
12m 46s | |
6m 50s | |
18m 2s | |
10m 25s | |
17m 34s | |
21m 8s | |
11m 28s | |
3h 0m 5s |