Training Issue
Ashwin-Ramesh2607 opened this issue · comments
Hey, I used the training command in ./train_1st_phase.sh with only one main change (removed the distributed training arguments since I have just 1 GPU). The training runs fine for 1 epoch but gets stuck during the test stage forever. (Testing stage since the total batches per epoch is 118)
Hi @Ashwin-Ramesh2607
Did you solve the problem? I'll check it soon. Thanks for reporting the bug, although I have tested carefully the source code.