unlockable/MediaNCognition
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Correct image normalization

Browse Source
This commit is contained in:
unlockable
2024-04-09 21:42:26 +08:00
parent 251a7e599a
commit 3747678e61
3 changed files with 6 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
.gitignore vendored
View File

@@ -7,4 +7,5 @@ __pycache__/
*.synctex.gz(buzy) *.synctex.gz(buzy)
*.out *.out
*.pdf *.pdf
.DS_Store .DS_Store
hw2/code/checkpoints/

4
hw2/code/datasets.py
View File

@@ -44,7 +44,7 @@ def get_data_loader(
transforms.Resize(image_size), transforms.Resize(image_size),
transforms.ToImage(), transforms.ToImage(),
transforms.ToDtype(torch.float32, scale=True), transforms.ToDtype(torch.float32, scale=True),
transforms.Normalize(mean=[-127.0, -127.0, -127.0], std=[128.0, 128.0, 128.0]) transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
] ]
# You should insert some data augmentation techniques to `data_transforms` when `augment` is True # You should insert some data augmentation techniques to `data_transforms` when `augment` is True
@@ -58,8 +58,6 @@ def get_data_loader(
# Use `transforms.Compose` to compose the list of transforms into a single transform # Use `transforms.Compose` to compose the list of transforms into a single transform
data_transforms = transforms.Compose(data_transforms) data_transforms = transforms.Compose(data_transforms)
print(type(data_transforms))
# >>> TODO 1.2: Define the dataset. # >>> TODO 1.2: Define the dataset.
# You should build the path to the selected dataset according to the `mode` parameter, # You should build the path to the selected dataset according to the `mode` parameter,
# and use the `ImageFolder` class from `torchvision.datasets` to load the datasets. # and use the `ImageFolder` class from `torchvision.datasets` to load the datasets.

6
hw2/code/networks.py
View File

@@ -187,7 +187,7 @@ class Classifier(nn.Module):
# Step 3: use `Tensor.view()` to flatten the tensor to match the size of the input of the # Step 3: use `Tensor.view()` to flatten the tensor to match the size of the input of the
# fully connected layers. # fully connected layers.
x = x.view(-1, 2048) x = x.view(x.shape[0], -1)
# Step 4: forward process for the fully connected network # Step 4: forward process for the fully connected network
out = self.fc_net(x) out = self.fc_net(x)
@@ -241,8 +241,8 @@ class STN(nn.Module):
# Suggested structure: 2 linear layers with one BN and ReLU. # Suggested structure: 2 linear layers with one BN and ReLU.
self.localization_fc = nn.Sequential( self.localization_fc = nn.Sequential(
nn.Linear(16, 256), nn.Linear(16, 256),
nn.Linear(256, 360), nn.Linear(256, 6),
nn.BatchNorm1d(360), nn.BatchNorm1d(6),
nn.ReLU() nn.ReLU()
) )
# <<< TODO 4.1 # <<< TODO 4.1