Skip to content

Tweak: New tqdm bars #13

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
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Tweak: New tqdm bars #13

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
12 changes: 5 additions & 7 deletions rf.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -95,7 +95,7 @@ def sample(self, z, cond, null_cond=None, sample_steps=50, cfg=2.0):
).cuda()

model_size = sum(p.numel() for p in model.parameters() if p.requires_grad)
print(f"Number of parameters: {model_size}, {model_size / 1e6}M")
print(f"Number of trainable parameters: {model_size}, {model_size / 1e6}M")

rf = RF(model)
optimizer = optim.Adam(model.parameters(), lr=5e-4)
Expand All @@ -106,10 +106,10 @@ def sample(self, z, cond, null_cond=None, sample_steps=50, cfg=2.0):

wandb.init(project=f"rf_{dataset_name}")

for epoch in range(100):
for epoch in tqdm(range(100), unit="epoch", leave=False, position=0):
lossbin = {i: 0 for i in range(10)}
losscnt = {i: 1e-6 for i in range(10)}
for i, (x, c) in tqdm(enumerate(dataloader)):
for i, (x, c) in tqdm(enumerate(dataloader), total=len(dataloader), unit="batch", leave=False, position=1):
x, c = x.cuda(), c.cuda()
optimizer.zero_grad()
loss, blsct = rf.forward(x, c)
Expand All @@ -123,10 +123,8 @@ def sample(self, z, cond, null_cond=None, sample_steps=50, cfg=2.0):
lossbin[int(t * 10)] += l
losscnt[int(t * 10)] += 1

# log
for i in range(10):
print(f"Epoch: {epoch}, {i} range loss: {lossbin[i] / losscnt[i]}")

time_bin_losses = [f"{i}:{lossbin[i]/losscnt[i]:.4f}" for i in range(10)]
tqdm.write(f"Epoch {epoch} time bin losses: {' '.join(time_bin_losses)}")
wandb.log({f"lossbin_{i}": lossbin[i] / losscnt[i] for i in range(10)})

rf.model.eval()
Expand Down