search

pbaylies / stylegan-encoder

StyleGAN Encoder - converts real images to latent space
Other
740 stars 180 forks source link

By default, resume training from latest snapshot #19

Closed shawwn closed 5 years ago

shawwn commented 5 years ago

Is this the correct place to set this? If so, would it be a good idea to resume from latest snapshot by default?

shawwn commented 5 years ago

FWIW, I just tested this and it works great.

The definition of locate_latest_pkl came from https://www.gwern.net/Faces#cb20-9