It can be continued to train?

mit-han-lab / gan-compression

[CVPR 2020] GAN Compression: Efficient Architectures for Interactive Conditional GANs

Other

1.1k stars 148 forks source link

It can be continued to train? #82

Closed PatrickHaopc closed 3 years ago

PatrickHaopc commented 3 years ago

It occurred to power off when I was training the MobileNet Teacher Model, can I continue to train from where it breaked off ?

lmxyy commented 3 years ago

Of course. You could use restore_G_path and restore_D_path to restore the generator and discriminator, and use restore_O_path to restore your optimizer. Also remember to change the epoch_base and iter_base to the correct value.