mxfont
mxfont copied to clipboard
font_indice font_indice
Excuse me, Where are font_indice, font_indice parameter passed from discriminator.py and what does it mean
class CustomDiscriminator(nn.Module): """ spectral norm + ResBlock + Multi-task Discriminator (No patchGAN) """ def init(self, feats, gap, projD): super().init() self.feats = feats self.gap = gap self.projD = projD def forward(self, x, font_indice, char_indice, out_feats='none'): assert out_feats in {'none', 'all'} feats = [] for layer in self.feats: x = layer(x) feats.append(x)
x = self.gap(x) # final features
ret = self.projD(x, font_indice, char_indice)
It is the label of font style, which is used for computing the adversarial loss. You can find the detailed explanation of the adversarial loss in Appendix B.4 of our paper.
thanks very much ! another question:The training set consists of content and style, with 6763 characters for content and 556 characters for style. May I ask if this dataset combination is reasonable