Skip to content

修复BCEWithLogitsLoss类型不匹配及save_image API兼容性问题 #40

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Dingchunru wants to merge 2 commits into
base: master
Choose a base branch
from
Open

修复BCEWithLogitsLoss类型不匹配及save_image API兼容性问题 #40

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f6c9719
Update utils.py
Dingchunru Jun 19, 2025
f71a674
Update hidden.py
Dingchunru Jun 19, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 7 additions & 8 deletions model/hidden.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -66,12 +66,11 @@ def train_on_batch(self, batch: list):
# ---------------- Train the discriminator -----------------------------
self.optimizer_discrim.zero_grad()
# train on cover
d_target_label_cover = torch.full((batch_size, 1), self.cover_label, device=self.device)
d_target_label_encoded = torch.full((batch_size, 1), self.encoded_label, device=self.device)
g_target_label_encoded = torch.full((batch_size, 1), self.cover_label, device=self.device)

d_target_label_cover = torch.full((batch_size, 1), self.cover_label, device=self.device, dtype=torch.float32)
d_target_label_encoded = torch.full((batch_size, 1), self.encoded_label, device=self.device, dtype=torch.float32)
g_target_label_encoded = torch.full((batch_size, 1), self.cover_label, device=self.device, dtype=torch.float32)
d_on_cover = self.discriminator(images)
d_loss_on_cover = self.bce_with_logits_loss(d_on_cover, d_target_label_cover)
d_loss_on_cover = self.bce_with_logits_loss(d_on_cover, d_target_label_cover.float())
d_loss_on_cover.backward()

# train on fake
Expand Down Expand Up @@ -144,15 +143,15 @@ def validate_on_batch(self, batch: list):
g_target_label_encoded = torch.full((batch_size, 1), self.cover_label, device=self.device)

d_on_cover = self.discriminator(images)
d_loss_on_cover = self.bce_with_logits_loss(d_on_cover, d_target_label_cover)
d_loss_on_cover = self.bce_with_logits_loss(d_on_cover, d_target_label_cover.float())

encoded_images, noised_images, decoded_messages = self.encoder_decoder(images, messages)

d_on_encoded = self.discriminator(encoded_images)
d_loss_on_encoded = self.bce_with_logits_loss(d_on_encoded, d_target_label_encoded)
d_loss_on_encoded = self.bce_with_logits_loss(d_on_encoded, d_target_label_encoded.float())

d_on_encoded_for_enc = self.discriminator(encoded_images)
g_loss_adv = self.bce_with_logits_loss(d_on_encoded_for_enc, g_target_label_encoded)
g_loss_adv = self.bce_with_logits_loss(d_on_encoded_for_enc, g_target_label_encoded.float())

if self.vgg_loss is None:
g_loss_enc = self.mse_loss(encoded_images, images)
Expand Down
12 changes: 10 additions & 2 deletions utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -54,7 +54,15 @@ def save_images(original_images, watermarked_images, epoch, folder, resize_to=No

stacked_images = torch.cat([images, watermarked_images], dim=0)
filename = os.path.join(folder, 'epoch-{}.png'.format(epoch))
torchvision.utils.save_image(stacked_images, filename, original_images.shape[0], normalize=False)

# 修复后的save_image调用
torchvision.utils.save_image(
stacked_images,
filename,
nrow=int(original_images.shape[0]), # 每行显示的图像数量
normalize=False,
format='png' # 明确指定格式
)


def sorted_nicely(l):
Expand Down Expand Up @@ -181,4 +189,4 @@ def write_losses(file_name, losses_accu, epoch, duration):
writer.writerow(row_to_write)
row_to_write = [epoch] + ['{:.4f}'.format(loss_avg.avg) for loss_avg in losses_accu.values()] + [
'{:.0f}'.format(duration)]
writer.writerow(row_to_write)
writer.writerow(row_to_write)