Skip to content

FIX Unet for SR tasks #8

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

FIX Unet for SR tasks #8

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
39 changes: 36 additions & 3 deletions solvers/u-net.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,26 @@
import deepinv as dinv


class UnetSR(torch.nn.Module):
def __init__(self, scale_factor):
super(UnetSR, self).__init__()
self.scale_factor = scale_factor
self.unet = dinv.models.UNet(
in_channels=3, out_channels=3*scale_factor**2,
scales=3, batch_norm=False
)
self.pixel_shuffle = torch.nn.PixelShuffle(scale_factor)
self.conv = torch.nn.Conv2d(
3, 3, kernel_size=3, padding=1, stride=1
)

def forward(self, x_input, physics, **kwargs):
x = self.unet(x_input, physics, **kwargs)
x = self.pixel_shuffle(x)
x = self.conv(x)
return x


class Solver(BaseSolver):

name = 'UNet'
Expand All @@ -24,6 +44,16 @@ def set_objective(self, train_dataset, physics):
self.train_dataloader = DataLoader(
train_dataset, batch_size=batch_size, shuffle=False
)

# Determine scale factor for super-resolution tasks
hr, lr = next(iter(self.train_dataloader))
_, _, h_lr, _ = lr.shape
_, _, h_hr, _ = hr.shape
if h_lr != h_hr:
self.scale_factor = h_hr // h_lr
else:
self.scale_factor = None

self.device = (
dinv.utils.get_freer_gpu() if torch.cuda.is_available() else "cpu"
)
Expand All @@ -32,9 +62,12 @@ def set_objective(self, train_dataset, physics):
def run(self, n_iter):
epochs = 4

model = dinv.models.UNet(
in_channels=3, out_channels=3, scales=3, batch_norm=False
).to(self.device)
if self.scale_factor is not None:
model = UnetSR(self.scale_factor).to(self.device)
else:
model = dinv.models.UNet(
in_channels=3, out_channels=3, scales=3, batch_norm=False
).to(self.device)

verbose = True # print training information
wandb_vis = False # plot curves and images in Weight&Bias
Expand Down
Loading