You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, congrats on the great work. I'm training this work in a 3090 GPU. However, as the number of training epochs is 1500, I need to train for nearly a month. Hence, I feel confused about the training duration。
The text was updated successfully, but these errors were encountered:
Hi, sorry for the confusion. The training only takes 150~250 epochs, not 1500. You can check if it converges or not according to the validation loss. As long as the validation loss doesn't decrease anymore, you can stop it.
Hi, congrats on the great work. I'm training this work in a 3090 GPU. However, as the number of training epochs is 1500, I need to train for nearly a month. Hence, I feel confused about the training duration。
The text was updated successfully, but these errors were encountered: