How to resume model training with train_supernet.sh?

mit-han-lab / gan-compression

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

Other

1.1k stars 148 forks source link

Closed YSL0226 closed 3 years ago

YSL0226 commented 3 years ago

My server was shutdown when I run train_supernet.sh a while. Is it possible to resume training? I took a look of parameters, but I didn't find it.

lmxyy commented 3 years ago

You could refer to #82. I think the resummation process is almost the same.