Closed jc-ryan closed 2 years ago
I was unable to reproduce this problem for a long time but now I have finally discovered the reason 😅 . It is caused by the difference in the pytorch-lightning
version.
See this https://pytorch-lightning.readthedocs.io/en/latest/common/trainer.html#checkpoint-callback for details, but the key is that the checkpoint_callback
should be passed in the Trainer
as part of the list of callbacks instead of being an individual argument.
Hello Sha Li, according to your code,I found that the checkpoint directory is still empty after the execution of code,which made it impossible to test the model on test set. Have you ever met this problem and could you give me some clue to fix it?