deviandice/DL-Art-School
1
0
Fork 0
forked from mrq/DL-Art-School
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Fix feature disc log item error

Browse Source
This commit is contained in:
James Betker 2020-07-26 22:25:59 -06:00
parent 9a8f227501
commit c54784ae9e
1 changed files with 5 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
codes/models/SRGAN_model.py
View File

@ -540,16 +540,16 @@ class SRGANModel(BaseModel):
if self.l_gan_w > 0: if self.l_gan_w > 0:
self.add_log_entry('l_g_gan', l_g_gan_log.item()) self.add_log_entry('l_g_gan', l_g_gan_log.item())
self.add_log_entry('l_g_total', l_g_total_log.item()) self.add_log_entry('l_g_total', l_g_total_log.item())
if self.l_gan_w > 0 and step > self.G_warmup:
self.add_log_entry('l_d_real', l_d_real_log.item())
self.add_log_entry('l_d_fake', l_d_fake_log.item())
self.add_log_entry('D_fake', torch.mean(pred_d_fake.detach()))
self.add_log_entry('D_diff', torch.mean(pred_d_fake) - torch.mean(pred_d_real))
if self.opt['train']['gan_type'] == 'pixgan_fea': if self.opt['train']['gan_type'] == 'pixgan_fea':
self.add_log_entry('l_d_fea_fake', l_d_fea_fake.item() * self.mega_batch_factor) self.add_log_entry('l_d_fea_fake', l_d_fea_fake.item() * self.mega_batch_factor)
self.add_log_entry('l_d_fea_real', l_d_fea_real.item() * self.mega_batch_factor) self.add_log_entry('l_d_fea_real', l_d_fea_real.item() * self.mega_batch_factor)
self.add_log_entry('l_d_fake_total', l_d_fake.item() * self.mega_batch_factor) self.add_log_entry('l_d_fake_total', l_d_fake.item() * self.mega_batch_factor)
self.add_log_entry('l_d_real_total', l_d_real.item() * self.mega_batch_factor) self.add_log_entry('l_d_real_total', l_d_real.item() * self.mega_batch_factor)
if self.l_gan_w > 0 and step > self.G_warmup:
self.add_log_entry('l_d_real', l_d_real_log.item())
self.add_log_entry('l_d_fake', l_d_fake_log.item())
self.add_log_entry('D_fake', torch.mean(pred_d_fake.detach()))
self.add_log_entry('D_diff', torch.mean(pred_d_fake) - torch.mean(pred_d_real))
# Log learning rates. # Log learning rates.
for i, pg in enumerate(self.optimizer_G.param_groups): for i, pg in enumerate(self.optimizer_G.param_groups):