You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Not really an issue, I'm just puzzled where max_size is used for. It serves as the size of an ImagePool pool and is used to store 'fake outputs'. Used here:
` # Update G network and record fake outputs
fake_A, fake_B, _, summary_str = self.sess.run(
[self.fake_A, self.fake_B, self.g_optim, self.g_sum],
feed_dict={self.real_data: batch_images, self.lr: lr})
self.writer.add_summary(summary_str, counter)
[fake_A, fake_B] = self.pool([fake_A, fake_B])
Not really an issue, I'm just puzzled where max_size is used for. It serves as the size of an ImagePool pool and is used to store 'fake outputs'. Used here:
` # Update G network and record fake outputs
fake_A, fake_B, _, summary_str = self.sess.run(
[self.fake_A, self.fake_B, self.g_optim, self.g_sum],
feed_dict={self.real_data: batch_images, self.lr: lr})
self.writer.add_summary(summary_str, counter)
[fake_A, fake_B] = self.pool([fake_A, fake_B])
`
Does the size influence training? Default it is set to 50. Any ideas on this?
Thanks!
The text was updated successfully, but these errors were encountered: