Saving the best model only during training process

affinelayer / pix2pix-tensorflow

Tensorflow port of Image-to-Image Translation with Conditional Adversarial Nets https://phillipi.github.io/pix2pix/

MIT License

5.08k stars 1.3k forks source link

Saving the best model only during training process #197

Open burhr2 opened 4 years ago

burhr2 commented 4 years ago

Hello guys, when specifying how frequent to save the model (--save_freq) in pix2pix.py does it save the best model or the current model?. My observation is that it saves the current model. If my observation is correct, how can I save the best model only during the training process? I using TensorFlow 1.14.0 as it seems to work fine in my case.