Pytorch-cyclegan-and-pix2pix: continue train

Created on 10 Aug 2020  路  1Comment  路  Source: junyanz/pytorch-CycleGAN-and-pix2pix

i have read all the issues,but still confused at the continue training. I have trained 200 epoch ,and i tested,i think the result is not good.i want to continue train,i set "python train.py --continue_train --epoch_count 200" ,while it just trained at 200th epoch and stoped.i want to train it to another 200 epochs,could you tell me how to set the parameters

Most helpful comment

You probably also need to change --n_epochs and --n_epochs_decay. (e.g., --n_epochs 200 --n_epochs_decay 200 --epoch_count 200 --continue_train.) See this line.

>All comments

You probably also need to change --n_epochs and --n_epochs_decay. (e.g., --n_epochs 200 --n_epochs_decay 200 --epoch_count 200 --continue_train.) See this line.

Was this page helpful?
0 / 5 - 0 ratings