hugo-xie
Results
1
comments of
hugo-xie
> Yes it's normal So does the loss remain in a fluctuating state from the beginning to the end of training? How do we then choose the final model checkpoint