Why training discriminator uses rec_img_small, in addition to decoder_big and decoder_part?

odegeasslbc / FastGAN-pytorch

Official implementation of the paper "Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis" in ICLR 2021

GNU General Public License v3.0

600 stars 100 forks source link

Why training discriminator uses rec_img_small, in addition to decoder_big and decoder_part? #32

Closed xuewengeophysics closed 3 years ago

xuewengeophysics commented 3 years ago

Hi. Thanks for your excellent research work. I have a question why training discriminator uses decoder_small, in addition to decoder_big and decoder_part? In my opinion, reconstruction loss does not use rec_img_small accroding to paper. Maybe I missed important information.

rec_img_big = self.decoder_big(feat_last) rec_img_small = self.decoder_small(feat_small) rec_img_part = self.decoder_part(feat_32[:,:,:8,:8])

odegeasslbc commented 3 years ago

Yeah, this "decoder_small" is an optional part, I just wanted to experiment on different structure configurations.