From 181b917656dfdf2e7a4cb9069bfed2de481c86ee Mon Sep 17 00:00:00 2001
From: Aavishkar04 <101863152+Aavishkar04@users.noreply.github.com>
Date: Sun, 13 Oct 2024 23:19:07 +0530
Subject: [PATCH] solved the issue

Resolved the issue related to deprecated methods in the torchvision library, specifically the use of transforms.RandomSizedCrop and transforms.Scale. These methods are outdated and should be replaced with transforms.RandomResizedCrop and transforms.Resize, respectively.

Solution:

Updated the data transformations to replace deprecated methods with their recommended alternatives.
Changes:

Replaced transforms.RandomSizedCrop(224) with transforms.RandomResizedCrop(224).
Replaced transforms.Scale(256) with transforms.Resize(256).
This resolves the warnings regarding deprecated transforms and ensures the code adheres to the latest version of the torchvision library.
---
 level1/train.py | 22 +++++++++++-----------
 1 file changed, 11 insertions(+), 11 deletions(-)

diff --git a/level1/train.py b/level1/train.py
index e2c729e..29565f5 100644
--- a/level1/train.py
+++ b/level1/train.py
@@ -32,7 +32,7 @@ def train_model(model, criterion, optimizer, scheduler, num_epochs=25):
             running_corrects = 0.0
 
             # Iterate over data.
-            for data in dataloders[phase]:
+            for data in dataloaders[phase]:  # Corrected the spelling from 'dataloders' to 'dataloaders'
                 # get the inputs
                 inputs, labels = data
 
@@ -57,7 +57,7 @@ def train_model(model, criterion, optimizer, scheduler, num_epochs=25):
                     optimizer.step()
 
                 # statistics
-                running_loss += loss.data[0]
+                running_loss += loss.item()  # Changed to .item() for correct loss extraction
                 running_corrects += torch.sum(preds == labels.data).to(torch.float32)
 
             epoch_loss = running_loss / dataset_sizes[phase]
@@ -82,38 +82,38 @@ def train_model(model, criterion, optimizer, scheduler, num_epochs=25):
 
 if __name__ == '__main__':
 
-    # data_transform, pay attention that the input of Normalize() is Tensor and the input of RandomResizedCrop() or RandomHorizontalFlip() is PIL Image
+    # Data transformations
     data_transforms = {
         'train': transforms.Compose([
-            transforms.RandomSizedCrop(224),
+            transforms.RandomResizedCrop(224),  # Changed from RandomSizedCrop to RandomResizedCrop
             transforms.RandomHorizontalFlip(),
             transforms.ToTensor(),
             transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
         ]),
         'val': transforms.Compose([
-            transforms.Scale(256),
+            transforms.Resize(256),  # Changed from Scale to Resize
             transforms.CenterCrop(224),
             transforms.ToTensor(),
             transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
         ]),
     }
 
-    # your image data file
+    # Your image data file
     data_dir = '/data'
     image_datasets = {x: datasets.ImageFolder(os.path.join(data_dir, x),
                                               data_transforms[x]) for x in ['train', 'val']}
-    # wrap your data and label into Tensor
-    dataloders = {x: torch.utils.data.DataLoader(image_datasets[x],
+    # Wrap your data and label into Tensor
+    dataloaders = {x: torch.utils.data.DataLoader(image_datasets[x],
                                                  batch_size=4,
                                                  shuffle=True,
                                                  num_workers=4) for x in ['train', 'val']}
 
     dataset_sizes = {x: len(image_datasets[x]) for x in ['train', 'val']}
 
-    # use gpu or not
+    # Use GPU or not
     use_gpu = torch.cuda.is_available()
 
-    # get model and replace the original fc layer with your fc layer
+    # Get model and replace the original fc layer with your fc layer
     model_ft = models.resnet18(pretrained=True)
     num_ftrs = model_ft.fc.in_features
     model_ft.fc = nn.Linear(num_ftrs, 2)
@@ -121,7 +121,7 @@ def train_model(model, criterion, optimizer, scheduler, num_epochs=25):
     if use_gpu:
         model_ft = model_ft.cuda()
 
-    # define loss function
+    # Define loss function
     criterion = nn.CrossEntropyLoss()
 
     # Observe that all parameters are being optimized