Question of finetuning in fast gan compression

mit-han-lab / gan-compression

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

Other

1.1k stars 148 forks source link

Question of finetuning in fast gan compression #86

Closed colorjam closed 2 years ago

colorjam commented 3 years ago

The process of finetuning is the same as supernet training, why specify the config_str?

lmxyy commented 2 years ago

The config_str is used to specify the subnet for finetuning. You do not need to sample a configuration for each iteration. Just fixing the specified subnet configuration is okay.