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 @GuGuLL123 , the epoch number is determined by the number of labeled images in USRN while by the number of all training images in CAC. To ensure the same number of training iterations, USRN increases the epoch number defined in CAC.
hello !
in main.py,
number_epochs = config['trainer']['epochs']
config['trainer']['epochs'] = int(config['num_images_all'] / config['n_labeled_examples']) * number_epochs
it changes the epoch number of trainer from 80 to 5120. Does the training process need such a large epoch number? or something wrong?
The text was updated successfully, but these errors were encountered: