pixel-recursive-super-resolution
pixel-recursive-super-resolution copied to clipboard
Load a model and resume training?
Is there a way to load the model and resume training? I am training on approx 75k images and it is taking almost 2 days (GTX 1080). So is there any way to resume training from the last checkpoint?
https://github.com/nilboy/pixel-recursive-super-resolution/blob/master/solver.py#L55 saver.restore(sess, './models/model.ckpt-30000')
@nilboy I am trying to load the last checkpoint files I have, in a replica file of solver.py and commented out the portion for loss, since I don't need backprop now and only calling the 'sample' function to save the image as per the session loaded.
Is this the right way to create an inference? If you have time, please guide me on how to create an inference. Or best, if you can upload a file for testing the trained models.