Fix PyTorchBackend
TP vs DP inputs distribution across replicas and shards
#584
Job | Run time |
---|---|
3m 3s | |
3m 9s | |
3m 3s | |
9m 15s |
PyTorchBackend
TP vs DP inputs distribution across replicas and shards
#584
Job | Run time |
---|---|
3m 3s | |
3m 9s | |
3m 3s | |
9m 15s |