TD2 Deep Learning.ipynb

 %% Cell type:markdown id:7edf7168 tags:
 
 # TD2: Deep learning
 
 %% Cell type:markdown id:fbb8c8df tags:
 
 In this TD, you must modify this notebook to answer the questions. To do this,
 
 1. Fork this repository
 2. Clone your forked repository on your local computer
 3. Answer the questions
 4. Commit and push regularly
 
 The last commit is due on Sunday, December 1, 11:59 PM. Later commits will not be taken into account.
 
 %% Cell type:markdown id:3d167a29 tags:
 
 Install and test PyTorch from  https://pytorch.org/get-started/locally.
 
 %% Cell type:code id:330a42f5 tags:
 
 ``` python
 %pip install torch torchvision
 ```
 
+%% Output
+
+    Requirement already satisfied: torch in c:\users\lucil\anaconda3\lib\site-packages (2.1.0)Note: you may need to restart the kernel to use updated packages.
+    
+    Requirement already satisfied: torchvision in c:\users\lucil\anaconda3\lib\site-packages (0.16.0)
+    Requirement already satisfied: filelock in c:\users\lucil\anaconda3\lib\site-packages (from torch) (3.9.0)
+    Requirement already satisfied: typing-extensions in c:\users\lucil\anaconda3\lib\site-packages (from torch) (4.7.1)
+    Requirement already satisfied: sympy in c:\users\lucil\anaconda3\lib\site-packages (from torch) (1.11.1)
+    Requirement already satisfied: networkx in c:\users\lucil\anaconda3\lib\site-packages (from torch) (3.1)
+    Requirement already satisfied: jinja2 in c:\users\lucil\anaconda3\lib\site-packages (from torch) (3.1.2)
+    Requirement already satisfied: fsspec in c:\users\lucil\anaconda3\lib\site-packages (from torch) (2023.4.0)
+    Requirement already satisfied: numpy in c:\users\lucil\anaconda3\lib\site-packages (from torchvision) (1.24.3)
+    Requirement already satisfied: requests in c:\users\lucil\anaconda3\lib\site-packages (from torchvision) (2.31.0)
+    Requirement already satisfied: pillow!=8.3.*,>=5.3.0 in c:\users\lucil\anaconda3\lib\site-packages (from torchvision) (9.4.0)
+    Requirement already satisfied: MarkupSafe>=2.0 in c:\users\lucil\anaconda3\lib\site-packages (from jinja2->torch) (2.1.1)
+    Requirement already satisfied: charset-normalizer<4,>=2 in c:\users\lucil\anaconda3\lib\site-packages (from requests->torchvision) (2.0.4)
+    Requirement already satisfied: idna<4,>=2.5 in c:\users\lucil\anaconda3\lib\site-packages (from requests->torchvision) (3.4)
+    Requirement already satisfied: urllib3<3,>=1.21.1 in c:\users\lucil\anaconda3\lib\site-packages (from requests->torchvision) (1.26.16)
+    Requirement already satisfied: certifi>=2017.4.17 in c:\users\lucil\anaconda3\lib\site-packages (from requests->torchvision) (2023.7.22)
+    Requirement already satisfied: mpmath>=0.19 in c:\users\lucil\anaconda3\lib\site-packages (from sympy->torch) (1.3.0)
+
 %% Cell type:markdown id:0882a636 tags:
 
 
 To test run the following code
 
 %% Cell type:code id:b1950f0a tags:
 
 ``` python
 import torch
 
 N, D = 14, 10
 x = torch.randn(N, D).type(torch.FloatTensor)
 print(x)
 
 from torchvision import models
 
 alexnet = models.alexnet()
 print(alexnet)
 ```
 
+%% Output
+
+    tensor([[ 0.3653,  0.6776,  1.4290,  1.3045, -0.1440, -1.9016,  0.1427,  0.6754,
+              0.0791,  0.6423],
+            [-1.3009,  0.1227,  0.4001,  0.6688,  0.1672, -0.5949,  0.3957, -0.6071,
+             -0.7747,  0.6197],
+            [-0.7347, -1.5540,  2.3525,  0.1084,  0.1178,  0.5596,  0.6267,  2.1786,
+             -0.5310, -0.6559],
+            [ 0.6326, -1.0263,  0.3332, -0.1291,  0.1675, -0.1014,  1.3175,  0.3264,
+             -0.1400,  0.7431],
+            [ 0.4699,  0.9845, -1.4050,  1.1468,  0.7983,  1.0263, -1.6672,  0.1562,
+             -0.0875, -1.9664],
+            [-0.3761, -0.8523,  1.5731, -2.0885, -1.5779,  0.6759,  0.4770,  1.5133,
+             -1.4350, -0.5716],
+            [ 0.0985, -0.1337, -0.3850,  0.3503, -0.4130, -0.7820, -1.1305,  1.0061,
+              0.0298, -1.4626],
+            [-0.0387, -1.7999, -2.1245,  0.2555,  0.1214,  0.5655,  0.5005,  1.0409,
+              0.8113, -0.2322],
+            [ 2.1456,  0.3775,  0.8248,  0.8468,  0.8631, -0.0429, -1.5679, -0.6221,
+             -1.1605,  0.5963],
+            [ 0.1601,  0.2023, -0.9813,  0.1316,  0.1114, -1.8421,  0.6188, -0.3290,
+              0.6238,  0.3155],
+            [-0.3864, -0.5559,  0.4249, -1.0155, -0.9137,  0.1228, -0.3569,  1.1107,
+             -0.5542,  1.2470],
+            [-0.6112, -0.5138,  1.1420, -0.0729,  1.1220, -0.1792,  1.0880,  0.8450,
+              0.6158, -0.9575],
+            [ 0.9272,  0.1329,  0.4858, -0.5643, -0.1636, -0.2209,  0.9413,  0.1729,
+              0.4400,  0.2477],
+            [-0.2307,  2.0693,  0.0898,  1.8634,  0.1166,  0.2212,  0.9382, -0.6915,
+             -1.9567,  0.2097]])
+    AlexNet(
+      (features): Sequential(
+        (0): Conv2d(3, 64, kernel_size=(11, 11), stride=(4, 4), padding=(2, 2))
+        (1): ReLU(inplace=True)
+        (2): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)
+        (3): Conv2d(64, 192, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
+        (4): ReLU(inplace=True)
+        (5): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)
+        (6): Conv2d(192, 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+        (7): ReLU(inplace=True)
+        (8): Conv2d(384, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+        (9): ReLU(inplace=True)
+        (10): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+        (11): ReLU(inplace=True)
+        (12): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)
+      )
+      (avgpool): AdaptiveAvgPool2d(output_size=(6, 6))
+      (classifier): Sequential(
+        (0): Dropout(p=0.5, inplace=False)
+        (1): Linear(in_features=9216, out_features=4096, bias=True)
+        (2): ReLU(inplace=True)
+        (3): Dropout(p=0.5, inplace=False)
+        (4): Linear(in_features=4096, out_features=4096, bias=True)
+        (5): ReLU(inplace=True)
+        (6): Linear(in_features=4096, out_features=1000, bias=True)
+      )
+    )
+
 %% Cell type:markdown id:23f266da tags:
 
 ## Exercise 1: CNN on CIFAR10
 
 The goal is to apply a Convolutional Neural Net (CNN) model on the CIFAR10 image dataset and test the accuracy of the model on the basis of image classification. Compare the Accuracy VS the neural network implemented during TD1.
 
 Have a look at the following documentation to be familiar with PyTorch.
 
 https://pytorch.org/tutorials/beginner/pytorch_with_examples.html
 
 https://pytorch.org/tutorials/beginner/deep_learning_60min_blitz.html
 
 %% Cell type:markdown id:4ba1c82d tags:
 
 You can test if GPU is available on your machine and thus train on it to speed up the process
 
 %% Cell type:code id:6e18f2fd tags:
 
 ``` python
 import torch
 
 # check if CUDA is available
 train_on_gpu = torch.cuda.is_available()
 
 if not train_on_gpu:
     print("CUDA is not available.  Training on CPU ...")
 else:
     print("CUDA is available!  Training on GPU ...")
 ```
 
+%% Output
+
+    CUDA is not available.  Training on CPU ...
+
 %% Cell type:markdown id:5cf214eb tags:
 
 Next we load the CIFAR10 dataset
 
 %% Cell type:code id:462666a2 tags:
 
 ``` python
 import numpy as np
 from torchvision import datasets, transforms
 from torch.utils.data.sampler import SubsetRandomSampler
 
 # number of subprocesses to use for data loading
 num_workers = 0
 # how many samples per batch to load
 batch_size = 20
 # percentage of training set to use as validation
 valid_size = 0.2
 
 # convert data to a normalized torch.FloatTensor
 transform = transforms.Compose(
     [transforms.ToTensor(), transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))]
 )
 
 # choose the training and test datasets
 train_data = datasets.CIFAR10("data", train=True, download=True, transform=transform)
 test_data = datasets.CIFAR10("data", train=False, download=True, transform=transform)
 
 # obtain training indices that will be used for validation
 num_train = len(train_data)
 indices = list(range(num_train))
 np.random.shuffle(indices)
 split = int(np.floor(valid_size * num_train))
 train_idx, valid_idx = indices[split:], indices[:split]
 
 # define samplers for obtaining training and validation batches
 train_sampler = SubsetRandomSampler(train_idx)
 valid_sampler = SubsetRandomSampler(valid_idx)
 
 # prepare data loaders (combine dataset and sampler)
 train_loader = torch.utils.data.DataLoader(
     train_data, batch_size=batch_size, sampler=train_sampler, num_workers=num_workers
 )
 valid_loader = torch.utils.data.DataLoader(
     train_data, batch_size=batch_size, sampler=valid_sampler, num_workers=num_workers
 )
 test_loader = torch.utils.data.DataLoader(
     test_data, batch_size=batch_size, num_workers=num_workers
 )
 
 # specify the image classes
 classes = [
     "airplane",
     "automobile",
     "bird",
     "cat",
     "deer",
     "dog",
     "frog",
     "horse",
     "ship",
     "truck",
 ]
 ```
 
+%% Output
+
+    Files already downloaded and verified
+    Files already downloaded and verified
+
 %% Cell type:markdown id:58ec3903 tags:
 
 CNN definition (this one is an example)
 
 %% Cell type:code id:317bf070 tags:
 
 ``` python
 import torch.nn as nn
 import torch.nn.functional as F
 
 # define the CNN architecture
 
 
 class Net(nn.Module):
     def __init__(self):
         super(Net, self).__init__()
         self.conv1 = nn.Conv2d(3, 6, 5)
         self.pool = nn.MaxPool2d(2, 2)
         self.conv2 = nn.Conv2d(6, 16, 5)
         self.fc1 = nn.Linear(16 * 5 * 5, 120)
         self.fc2 = nn.Linear(120, 84)
         self.fc3 = nn.Linear(84, 10)
 
     def forward(self, x):
         x = self.pool(F.relu(self.conv1(x)))
         x = self.pool(F.relu(self.conv2(x)))
         x = x.view(-1, 16 * 5 * 5)
         x = F.relu(self.fc1(x))
         x = F.relu(self.fc2(x))
         x = self.fc3(x)
         return x
 
 
 # create a complete CNN
 model = Net()
 print(model)
 # move tensors to GPU if CUDA is available
 if train_on_gpu:
     model.cuda()
 ```
 
+%% Output
+
+    Net(
+      (conv1): Conv2d(3, 6, kernel_size=(5, 5), stride=(1, 1))
+      (pool): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
+      (conv2): Conv2d(6, 16, kernel_size=(5, 5), stride=(1, 1))
+      (fc1): Linear(in_features=400, out_features=120, bias=True)
+      (fc2): Linear(in_features=120, out_features=84, bias=True)
+      (fc3): Linear(in_features=84, out_features=10, bias=True)
+    )
+
 %% Cell type:markdown id:a2dc4974 tags:
 
 Loss function and training using SGD (Stochastic Gradient Descent) optimizer
 
 %% Cell type:code id:4b53f229 tags:
 
 ``` python
 import torch.optim as optim
 
 criterion = nn.CrossEntropyLoss()  # specify loss function
 optimizer = optim.SGD(model.parameters(), lr=0.01)  # specify optimizer
 
 n_epochs = 30  # number of epochs to train the model
 train_loss_list = []  # list to store loss to visualize
 valid_loss_min = np.Inf  # track change in validation loss
 
 for epoch in range(n_epochs):
     # Keep track of training and validation loss
     train_loss = 0.0
     valid_loss = 0.0
 
     # Train the model
     model.train()
     for data, target in train_loader:
         # Move tensors to GPU if CUDA is available
         if train_on_gpu:
             data, target = data.cuda(), target.cuda()
         # Clear the gradients of all optimized variables
         optimizer.zero_grad()
         # Forward pass: compute predicted outputs by passing inputs to the model
         output = model(data)
         # Calculate the batch loss
         loss = criterion(output, target)
         # Backward pass: compute gradient of the loss with respect to model parameters
         loss.backward()
         # Perform a single optimization step (parameter update)
         optimizer.step()
         # Update training loss
         train_loss += loss.item() * data.size(0)
 
     # Validate the model
     model.eval()
     for data, target in valid_loader:
         # Move tensors to GPU if CUDA is available
         if train_on_gpu:
             data, target = data.cuda(), target.cuda()
         # Forward pass: compute predicted outputs by passing inputs to the model
         output = model(data)
         # Calculate the batch loss
         loss = criterion(output, target)
         # Update average validation loss
         valid_loss += loss.item() * data.size(0)
 
     # Calculate average losses
     train_loss = train_loss / len(train_loader)
     valid_loss = valid_loss / len(valid_loader)
     train_loss_list.append(train_loss)
 
     # Print training/validation statistics
     print(
         "Epoch: {} \tTraining Loss: {:.6f} \tValidation Loss: {:.6f}".format(
             epoch, train_loss, valid_loss
         )
     )
 
     # Save model if validation loss has decreased
     if valid_loss <= valid_loss_min:
         print(
             "Validation loss decreased ({:.6f} --> {:.6f}).  Saving model ...".format(
                 valid_loss_min, valid_loss
             )
         )
         torch.save(model.state_dict(), "model_cifar.pt")
         valid_loss_min = valid_loss
 ```
 
+%% Output
+
+    Epoch: 0 	Training Loss: 43.453638 	Validation Loss: 38.117901
+    Validation loss decreased (inf --> 38.117901).  Saving model ...
+    Epoch: 1 	Training Loss: 33.786905 	Validation Loss: 30.608687
+    Validation loss decreased (38.117901 --> 30.608687).  Saving model ...
+    Epoch: 2 	Training Loss: 29.978750 	Validation Loss: 28.626190
+    Validation loss decreased (30.608687 --> 28.626190).  Saving model ...
+    Epoch: 3 	Training Loss: 27.777584 	Validation Loss: 27.198099
+    Validation loss decreased (28.626190 --> 27.198099).  Saving model ...
+    Epoch: 4 	Training Loss: 26.117933 	Validation Loss: 26.415911
+    Validation loss decreased (27.198099 --> 26.415911).  Saving model ...
+    Epoch: 5 	Training Loss: 24.786261 	Validation Loss: 24.554481
+    Validation loss decreased (26.415911 --> 24.554481).  Saving model ...
+    Epoch: 6 	Training Loss: 23.703873 	Validation Loss: 24.357461
+    Validation loss decreased (24.554481 --> 24.357461).  Saving model ...
+    Epoch: 7 	Training Loss: 22.748076 	Validation Loss: 24.332178
+    Validation loss decreased (24.357461 --> 24.332178).  Saving model ...
+    Epoch: 8 	Training Loss: 21.790853 	Validation Loss: 23.261406
+    Validation loss decreased (24.332178 --> 23.261406).  Saving model ...
+    Epoch: 9 	Training Loss: 20.925274 	Validation Loss: 23.353505
+    Epoch: 10 	Training Loss: 20.174014 	Validation Loss: 22.972180
+    Validation loss decreased (23.261406 --> 22.972180).  Saving model ...
+    Epoch: 11 	Training Loss: 19.419566 	Validation Loss: 22.647662
+    Validation loss decreased (22.972180 --> 22.647662).  Saving model ...
+    Epoch: 12 	Training Loss: 18.719525 	Validation Loss: 22.919457
+
+    ---------------------------------------------------------------------------
+    KeyboardInterrupt                         Traceback (most recent call last)
+    d:\Users\lucil\Documents\S9\Apprentissage profond\mod_4_6-td2\TD2 Deep Learning.ipynb Cell 15 line 3
+         <a href='vscode-notebook-cell:/d%3A/Users/lucil/Documents/S9/Apprentissage%20profond/mod_4_6-td2/TD2%20Deep%20Learning.ipynb#X20sZmlsZQ%3D%3D?line=33'>34</a> # Validate the model
+         <a href='vscode-notebook-cell:/d%3A/Users/lucil/Documents/S9/Apprentissage%20profond/mod_4_6-td2/TD2%20Deep%20Learning.ipynb#X20sZmlsZQ%3D%3D?line=34'>35</a> model.eval()
+    ---> <a href='vscode-notebook-cell:/d%3A/Users/lucil/Documents/S9/Apprentissage%20profond/mod_4_6-td2/TD2%20Deep%20Learning.ipynb#X20sZmlsZQ%3D%3D?line=35'>36</a> for data, target in valid_loader:
+         <a href='vscode-notebook-cell:/d%3A/Users/lucil/Documents/S9/Apprentissage%20profond/mod_4_6-td2/TD2%20Deep%20Learning.ipynb#X20sZmlsZQ%3D%3D?line=36'>37</a>     # Move tensors to GPU if CUDA is available
+         <a href='vscode-notebook-cell:/d%3A/Users/lucil/Documents/S9/Apprentissage%20profond/mod_4_6-td2/TD2%20Deep%20Learning.ipynb#X20sZmlsZQ%3D%3D?line=37'>38</a>     if train_on_gpu:
+         <a href='vscode-notebook-cell:/d%3A/Users/lucil/Documents/S9/Apprentissage%20profond/mod_4_6-td2/TD2%20Deep%20Learning.ipynb#X20sZmlsZQ%3D%3D?line=38'>39</a>         data, target = data.cuda(), target.cuda()
+File     c:\Users\lucil\anaconda3\Lib\site-packages\torch\utils\data\dataloader.py:630, in _BaseDataLoaderIter.__next__(self)
+        627 if self._sampler_iter is None:
+        628     # TODO(https://github.com/pytorch/pytorch/issues/76750)
+        629     self._reset()  # type: ignore[call-arg]
+    --> 630 data = self._next_data()
+        631 self._num_yielded += 1
+        632 if self._dataset_kind == _DatasetKind.Iterable and \
+        633         self._IterableDataset_len_called is not None and \
+        634         self._num_yielded > self._IterableDataset_len_called:
+File     c:\Users\lucil\anaconda3\Lib\site-packages\torch\utils\data\dataloader.py:674, in _SingleProcessDataLoaderIter._next_data(self)
+        672 def _next_data(self):
+        673     index = self._next_index()  # may raise StopIteration
+    --> 674     data = self._dataset_fetcher.fetch(index)  # may raise StopIteration
+        675     if self._pin_memory:
+        676         data = _utils.pin_memory.pin_memory(data, self._pin_memory_device)
+File     c:\Users\lucil\anaconda3\Lib\site-packages\torch\utils\data\_utils\fetch.py:51, in _MapDatasetFetcher.fetch(self, possibly_batched_index)
+         49         data = self.dataset.__getitems__(possibly_batched_index)
+         50     else:
+    ---> 51         data = [self.dataset[idx] for idx in possibly_batched_index]
+         52 else:
+         53     data = self.dataset[possibly_batched_index]
+File     c:\Users\lucil\anaconda3\Lib\site-packages\torch\utils\data\_utils\fetch.py:51, in <listcomp>(.0)
+         49         data = self.dataset.__getitems__(possibly_batched_index)
+         50     else:
+    ---> 51         data = [self.dataset[idx] for idx in possibly_batched_index]
+         52 else:
+         53     data = self.dataset[possibly_batched_index]
+File     c:\Users\lucil\anaconda3\Lib\site-packages\torchvision\datasets\cifar.py:118, in CIFAR10.__getitem__(self, index)
+        115 img = Image.fromarray(img)
+        117 if self.transform is not None:
+    --> 118     img = self.transform(img)
+        120 if self.target_transform is not None:
+        121     target = self.target_transform(target)
+File     c:\Users\lucil\anaconda3\Lib\site-packages\torchvision\transforms\transforms.py:95, in Compose.__call__(self, img)
+         93 def __call__(self, img):
+         94     for t in self.transforms:
+    ---> 95         img = t(img)
+         96     return img
+File     c:\Users\lucil\anaconda3\Lib\site-packages\torchvision\transforms\transforms.py:137, in ToTensor.__call__(self, pic)
+        129 def __call__(self, pic):
+        130     """
+        131     Args:
+        132         pic (PIL Image or numpy.ndarray): Image to be converted to tensor.
+       (...)
+        135         Tensor: Converted image.
+        136     """
+    --> 137     return F.to_tensor(pic)
+File     c:\Users\lucil\anaconda3\Lib\site-packages\torchvision\transforms\functional.py:174, in to_tensor(pic)
+        172 img = img.permute((2, 0, 1)).contiguous()
+        173 if isinstance(img, torch.ByteTensor):
+    --> 174     return img.to(dtype=default_float_dtype).div(255)
+        175 else:
+        176     return img
+    KeyboardInterrupt:
+
 %% Cell type:markdown id:13e1df74 tags:
 
 Does overfit occur? If so, do an early stopping.
 
 %% Cell type:code id:d39df818 tags:
 
 ``` python
 import matplotlib.pyplot as plt
 
-plt.plot(range(n_epochs), train_loss_list)
+n_epochs_overfit = 13 #Otherwise len(train_lost_list) < n_epochs
+plt.plot(range(n_epochs_overfit), train_loss_list)
 plt.xlabel("Epoch")
 plt.ylabel("Loss")
 plt.title("Performance of Model 1")
 plt.show()
 ```
 
+%% Output
+
+
+
 %% Cell type:markdown id:11df8fd4 tags:
 
 Now loading the model with the lowest validation loss value
 
 %% Cell type:code id:e93efdfc tags:
 
 ``` python
 model.load_state_dict(torch.load("./model_cifar.pt"))
 
 # track test loss
 test_loss = 0.0
 class_correct = list(0.0 for i in range(10))
 class_total = list(0.0 for i in range(10))
 
 model.eval()
 # iterate over test data
 for data, target in test_loader:
     # move tensors to GPU if CUDA is available
     if train_on_gpu:
         data, target = data.cuda(), target.cuda()
     # forward pass: compute predicted outputs by passing inputs to the model
     output = model(data)
     # calculate the batch loss
     loss = criterion(output, target)
     # update test loss
     test_loss += loss.item() * data.size(0)
     # convert output probabilities to predicted class
     _, pred = torch.max(output, 1)
     # compare predictions to true label
     correct_tensor = pred.eq(target.data.view_as(pred))
     correct = (
         np.squeeze(correct_tensor.numpy())
         if not train_on_gpu
         else np.squeeze(correct_tensor.cpu().numpy())
     )
     # calculate test accuracy for each object class
     for i in range(batch_size):
         label = target.data[i]
         class_correct[label] += correct[i].item()
         class_total[label] += 1
 
 # average test loss
 test_loss = test_loss / len(test_loader)
 print("Test Loss: {:.6f}\n".format(test_loss))
 
 for i in range(10):
     if class_total[i] > 0:
         print(
             "Test Accuracy of %5s: %2d%% (%2d/%2d)"
             % (
                 classes[i],
                 100 * class_correct[i] / class_total[i],
                 np.sum(class_correct[i]),
                 np.sum(class_total[i]),
             )
         )
     else:
         print("Test Accuracy of %5s: N/A (no training examples)" % (classes[i]))
 
 print(
     "\nTest Accuracy (Overall): %2d%% (%2d/%2d)"
     % (
         100.0 * np.sum(class_correct) / np.sum(class_total),
         np.sum(class_correct),
         np.sum(class_total),
     )
 )
 ```
 
+%% Output
+
+    Test Loss: 22.235297
+    
+    Test Accuracy of airplane: 52% (523/1000)
+    Test Accuracy of automobile: 84% (849/1000)
+    Test Accuracy of  bird: 34% (341/1000)
+    Test Accuracy of   cat: 43% (432/1000)
+    Test Accuracy of  deer: 66% (662/1000)
+    Test Accuracy of   dog: 44% (448/1000)
+    Test Accuracy of  frog: 74% (746/1000)
+    Test Accuracy of horse: 64% (647/1000)
+    Test Accuracy of  ship: 83% (836/1000)
+    Test Accuracy of truck: 64% (649/1000)
+    
+    Test Accuracy (Overall): 61% (6133/10000)
+
 %% Cell type:markdown id:944991a2 tags:
 
 Build a new network with the following structure.
 
 - It has 3 convolutional layers of kernel size 3 and padding of 1.
 - The first convolutional layer must output 16 channels, the second 32 and the third 64.
 - At each convolutional layer output, we apply a ReLU activation then a MaxPool with kernel size of 2.
 - Then, three fully connected layers, the first two being followed by a ReLU activation and a dropout whose value you will suggest.
 - The first fully connected layer will have an output size of 512.
 - The second fully connected layer will have an output size of 64.
 
 Compare the results obtained with this new network to those obtained previously.
 
+%% Cell type:markdown id: tags:
+
+CNN definition following the structure required above
+
+%% Cell type:code id: tags:
+
+``` python
+import torch.nn as nn
+import torch.nn.functional as F
+
+# define the CNN architecture
+
+
+class Net_1(nn.Module):
+    def __init__(self):
+        super(Net_1, self).__init__()
+
+        #Define the 3 convolutional layers
+        #First layer : 3 input channels, 16 output channels, kernel size 3, padding 1
+        self.conv1 = nn.Conv2d(in_channels=3, out_channels=16, kernel_size=3, padding=1)
+        #Second layer : 16 input channels, 32 output channels, kernel size 3, padding 1
+        self.conv2 = nn.Conv2d(in_channels=16, out_channels=32, kernel_size=3, padding=1)
+        #Third layer : 32 input channels, 64 output channels, kernel size 3, padding 1
+        self.conv3 = nn.Conv2d(in_channels=32, out_channels=64, kernel_size=3, padding=1)
+
+        #MaxPool with kernel size 2
+        self.pool = nn.MaxPool2d(2, 2)
+
+        #Define the 3 fully connected layers
+        #First layer : input of size 64, image dimension 4*4, output of size 512
+        self.fc1 = nn.Linear(in_features=64 * 4 * 4,out_features=512)
+        #Second layer : input of size 512, output of size 64
+        self.fc2 = nn.Linear(512, 64)
+        #Third layer : input of size 64, output of size 10
+        self.fc3 = nn.Linear(64, 10)
+
+        #Dropout
+        self.dropout = nn.Dropout(p=0.5)
+
+
+    def forward(self, x):
+
+        #Through the 3 convolutional layers
+        x = self.pool(F.relu(self.conv1(x)))
+        x = self.pool(F.relu(self.conv2(x)))
+        x = self.pool(F.relu(self.conv3(x)))
+
+        #Linearize the output
+        x = x.view(-1, 64 * 4 * 4)
+
+        #Through the 3 fully connected layers
+        x = self.dropout(F.relu(self.fc1(x)))
+        x = self.dropout(F.relu(self.fc2(x)))
+        x = self.fc3(x)
+
+        return x
+
+
+# create a complete CNN
+model_1 = Net_1()
+print(model_1)
+# move tensors to GPU if CUDA is available
+if train_on_gpu:
+    model.cuda()
+```
+
+%% Output
+
+    Net_1(
+      (conv1): Conv2d(3, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+      (conv2): Conv2d(16, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+      (conv3): Conv2d(32, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
+      (pool): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
+      (fc1): Linear(in_features=1024, out_features=512, bias=True)
+      (fc2): Linear(in_features=512, out_features=64, bias=True)
+      (fc3): Linear(in_features=64, out_features=10, bias=True)
+      (dropout): Dropout(p=0.5, inplace=False)
+    )
+
+%% Cell type:markdown id: tags:
+
+Loss function and training using SGD (Stochastic Gradient Descent) optimizer
+
+%% Cell type:code id: tags:
+
+``` python
+import torch.optim as optim
+
+criterion = nn.CrossEntropyLoss()  # specify loss function
+optimizer = optim.SGD(model_1.parameters(), lr=0.01)  # specify optimizer
+
+n_epochs_1 = 30  # number of epochs to train the model
+train_loss_list_1 = []  # list to store loss to visualize
+valid_loss_min_1 = np.Inf  # track change in validation loss
+
+for epoch in range(n_epochs_1):
+    # Keep track of training and validation loss
+    train_loss = 0.0
+    valid_loss = 0.0
+
+    # Train the model
+    model_1.train()
+    for data, target in train_loader:
+        # Move tensors to GPU if CUDA is available
+        if train_on_gpu:
+            data, target = data.cuda(), target.cuda()
+        # Clear the gradients of all optimized variables
+        optimizer.zero_grad()
+        # Forward pass: compute predicted outputs by passing inputs to the model
+        output = model_1(data)
+        # Calculate the batch loss
+        loss = criterion(output, target)
+        # Backward pass: compute gradient of the loss with respect to model parameters
+        loss.backward()
+        # Perform a single optimization step (parameter update)
+        optimizer.step()
+        # Update training loss
+        train_loss += loss.item() * data.size(0)
+
+    # Validate the model
+    model_1.eval()
+    for data, target in valid_loader:
+        # Move tensors to GPU if CUDA is available
+        if train_on_gpu:
+            data, target = data.cuda(), target.cuda()
+        # Forward pass: compute predicted outputs by passing inputs to the model
+        output = model_1(data)
+        # Calculate the batch loss
+        loss = criterion(output, target)
+        # Update average validation loss
+        valid_loss += loss.item() * data.size(0)
+
+    # Calculate average losses
+    train_loss = train_loss / len(train_loader)
+    valid_loss = valid_loss / len(valid_loader)
+    train_loss_list_1.append(train_loss)
+
+    # Print training/validation statistics
+    print(
+        "Epoch: {} \tTraining Loss: {:.6f} \tValidation Loss: {:.6f}".format(
+            epoch, train_loss, valid_loss
+        )
+    )
+
+    # Save model if validation loss has decreased
+    if valid_loss <= valid_loss_min_1:
+        print(
+            "Validation loss decreased ({:.6f} --> {:.6f}).  Saving model_1 ...".format(
+                valid_loss_min_1, valid_loss
+            )
+        )
+        torch.save(model_1.state_dict(), "model_1_cifar.pt")
+        valid_loss_min_1 = valid_loss
+```
+
+%% Output
+
+    Epoch: 0 	Training Loss: 45.348058 	Validation Loss: 41.718214
+    Validation loss decreased (inf --> 41.718214).  Saving model_1 ...
+    Epoch: 1 	Training Loss: 39.649087 	Validation Loss: 35.754235
+    Validation loss decreased (41.718214 --> 35.754235).  Saving model_1 ...
+    Epoch: 2 	Training Loss: 35.008029 	Validation Loss: 31.420939
+    Validation loss decreased (35.754235 --> 31.420939).  Saving model_1 ...
+    Epoch: 3 	Training Loss: 32.138094 	Validation Loss: 28.863286
+    Validation loss decreased (31.420939 --> 28.863286).  Saving model_1 ...
+    Epoch: 4 	Training Loss: 30.218731 	Validation Loss: 28.003921
+    Validation loss decreased (28.863286 --> 28.003921).  Saving model_1 ...
+    Epoch: 5 	Training Loss: 28.807953 	Validation Loss: 26.228902
+    Validation loss decreased (28.003921 --> 26.228902).  Saving model_1 ...
+    Epoch: 6 	Training Loss: 27.365782 	Validation Loss: 25.497843
+    Validation loss decreased (26.228902 --> 25.497843).  Saving model_1 ...
+    Epoch: 7 	Training Loss: 26.038266 	Validation Loss: 23.508494
+    Validation loss decreased (25.497843 --> 23.508494).  Saving model_1 ...
+    Epoch: 8 	Training Loss: 24.863525 	Validation Loss: 23.421283
+    Validation loss decreased (23.508494 --> 23.421283).  Saving model_1 ...
+    Epoch: 9 	Training Loss: 23.610995 	Validation Loss: 21.928674
+    Validation loss decreased (23.421283 --> 21.928674).  Saving model_1 ...
+    Epoch: 10 	Training Loss: 22.689530 	Validation Loss: 21.890606
+    Validation loss decreased (21.928674 --> 21.890606).  Saving model_1 ...
+    Epoch: 11 	Training Loss: 21.605674 	Validation Loss: 20.122198
+    Validation loss decreased (21.890606 --> 20.122198).  Saving model_1 ...
+    Epoch: 12 	Training Loss: 20.795100 	Validation Loss: 20.151628
+
+    ---------------------------------------------------------------------------
+    KeyboardInterrupt                         Traceback (most recent call last)
+    d:\Users\lucil\Documents\S9\Apprentissage profond\mod_4_6-td2\TD2 Deep Learning.ipynb Cell 24 line 2
+         <a href='vscode-notebook-cell:/d%3A/Users/lucil/Documents/S9/Apprentissage%20profond/mod_4_6-td2/TD2%20Deep%20Learning.ipynb#X56sZmlsZQ%3D%3D?line=25'>26</a> loss = criterion(output, target)
+         <a href='vscode-notebook-cell:/d%3A/Users/lucil/Documents/S9/Apprentissage%20profond/mod_4_6-td2/TD2%20Deep%20Learning.ipynb#X56sZmlsZQ%3D%3D?line=26'>27</a> # Backward pass: compute gradient of the loss with respect to model parameters
+    ---> <a href='vscode-notebook-cell:/d%3A/Users/lucil/Documents/S9/Apprentissage%20profond/mod_4_6-td2/TD2%20Deep%20Learning.ipynb#X56sZmlsZQ%3D%3D?line=27'>28</a> loss.backward()
+         <a href='vscode-notebook-cell:/d%3A/Users/lucil/Documents/S9/Apprentissage%20profond/mod_4_6-td2/TD2%20Deep%20Learning.ipynb#X56sZmlsZQ%3D%3D?line=28'>29</a> # Perform a single optimization step (parameter update)
+         <a href='vscode-notebook-cell:/d%3A/Users/lucil/Documents/S9/Apprentissage%20profond/mod_4_6-td2/TD2%20Deep%20Learning.ipynb#X56sZmlsZQ%3D%3D?line=29'>30</a> optimizer.step()
+File     c:\Users\lucil\anaconda3\Lib\site-packages\torch\_tensor.py:492, in Tensor.backward(self, gradient, retain_graph, create_graph, inputs)
+        482 if has_torch_function_unary(self):
+        483     return handle_torch_function(
+        484         Tensor.backward,
+        485         (self,),
+       (...)
+        490         inputs=inputs,
+        491     )
+    --> 492 torch.autograd.backward(
+        493     self, gradient, retain_graph, create_graph, inputs=inputs
+        494 )
+File     c:\Users\lucil\anaconda3\Lib\site-packages\torch\autograd\__init__.py:251, in backward(tensors, grad_tensors, retain_graph, create_graph, grad_variables, inputs)
+        246     retain_graph = create_graph
+        248 # The reason we repeat the same comment below is that
+        249 # some Python versions print out the first line of a multi-line function
+        250 # calls in the traceback and some print out the last line
+    --> 251 Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
+        252     tensors,
+        253     grad_tensors_,
+        254     retain_graph,
+        255     create_graph,
+        256     inputs,
+        257     allow_unreachable=True,
+        258     accumulate_grad=True,
+        259 )
+    KeyboardInterrupt:
+
+%% Cell type:markdown id: tags:
+
+Compare the results with the previous model's results
+
+%% Cell type:code id: tags:
+
+``` python
+import matplotlib.pyplot as plt
+
+n_epochs_overfit_1 = 13 #Otherwise len(train_lost_list) < n_epochs
+plt.plot(range(n_epochs_overfit), train_loss_list, label = "Model 0")
+plt.plot(range(n_epochs_overfit), train_loss_list_1, color = "green", label = "Model 1")
+plt.xlabel("Epoch")
+plt.ylabel("Loss")
+plt.title("Comparison of Performande for of Model 0 and Model 1")
+plt.show()
+```
+
+%% Output
+
+
+
+%% Cell type:code id: tags:
+
+``` python
+print(train_loss_list)
+print(train_loss_list_1)
+```
+
+%% Output
+
+    [43.45363824605942, 33.786904842853545, 29.978750259280204, 27.777584496736527, 26.11793281197548, 24.786260991096498, 23.703872640132904, 22.74807582437992, 21.79085268616676, 20.92527386188507, 20.174014331400393, 19.419565526545046, 18.71952503979206]
+    [45.348057844638824, 39.64908684611321, 35.00802879333496, 32.13809435069561, 30.21873086452484, 28.807953109145163, 27.365781868696214, 26.038266357183456, 24.863524509072302, 23.610995230078696, 22.689530485272407, 21.60567447721958, 20.795099827349187]
+
 %% Cell type:markdown id:bc381cf4 tags:
 
 ## Exercise 2: Quantization: try to compress the CNN to save space
 
 Quantization doc is available from https://pytorch.org/docs/stable/quantization.html#torch.quantization.quantize_dynamic
 
 The Exercise is to quantize post training the above CNN model. Compare the size reduction and the impact on the classification accuracy
 
 
 The size of the model is simply the size of the file.
 
 %% Cell type:code id:ef623c26 tags:
 
 ``` python
 import os
 
 
 def print_size_of_model(model, label=""):
     torch.save(model.state_dict(), "temp.p")
     size = os.path.getsize("temp.p")
     print("model: ", label, " \t", "Size (KB):", size / 1e3)
     os.remove("temp.p")
     return size
 
 
 print_size_of_model(model, "fp32")
 ```
 
+%% Output
+
+    model:  fp32  	 Size (KB): 251.278
+
+    251278
+
 %% Cell type:markdown id:05c4e9ad tags:
 
 Post training quantization example
 
 %% Cell type:code id:c4c65d4b tags:
 
 ``` python
 import torch.quantization
 
 
 quantized_model = torch.quantization.quantize_dynamic(model, dtype=torch.qint8)
 print_size_of_model(quantized_model, "int8")
 ```
 
+%% Output
+
+    model:  int8  	 Size (KB): 76.522
+
+    76522
+
 %% Cell type:markdown id:7b108e17 tags:
 
 For each class, compare the classification test accuracy of the initial model and the quantized model. Also give the overall test accuracy for both models.
 
 %% Cell type:markdown id:a0a34b90 tags:
 
 Try training aware quantization to mitigate the impact on the accuracy (doc available here https://pytorch.org/docs/stable/quantization.html#torch.quantization.quantize_dynamic)
 
+%% Cell type:markdown id: tags:
+
+First for the initial model :
+
+%% Cell type:code id: tags:
+
+``` python
+# import model
+model.load_state_dict(torch.load("./model_cifar.pt"))
+
+# track test loss
+test_loss = 0.0
+class_correct = list(0.0 for i in range(10))
+class_total = list(0.0 for i in range(10))
+
+model.eval()
+# iterate over test data
+for data, target in test_loader:
+    # move tensors to GPU if CUDA is available
+    if train_on_gpu:
+        data, target = data.cuda(), target.cuda()
+    # forward pass: compute predicted outputs by passing inputs to the model
+    output = model(data)
+    # calculate the batch loss
+    loss = criterion(output, target)
+    # update test loss
+    test_loss += loss.item() * data.size(0)
+    # convert output probabilities to predicted class
+    _, pred = torch.max(output, 1)
+    # compare predictions to true label
+    correct_tensor = pred.eq(target.data.view_as(pred))
+    correct = (
+        np.squeeze(correct_tensor.numpy())
+        if not train_on_gpu
+        else np.squeeze(correct_tensor.cpu().numpy())
+    )
+    # calculate test accuracy for each object class
+    for i in range(batch_size):
+        label = target.data[i]
+        class_correct[label] += correct[i].item()
+        class_total[label] += 1
+
+# average test loss
+test_loss = test_loss / len(test_loader)
+print("Test Loss: {:.6f}\n".format(test_loss))
+
+for i in range(10):
+    if class_total[i] > 0:
+        print(
+            "Test Accuracy of %5s: %2d%% (%2d/%2d)"
+            % (
+                classes[i],
+                100 * class_correct[i] / class_total[i],
+                np.sum(class_correct[i]),
+                np.sum(class_total[i]),
+            )
+        )
+    else:
+        print("Test Accuracy of %5s: N/A (no training examples)" % (classes[i]))
+
+print(
+    "\nTest Accuracy (Overall): %2d%% (%2d/%2d)"
+    % (
+        100.0 * np.sum(class_correct) / np.sum(class_total),
+        np.sum(class_correct),
+        np.sum(class_total),
+    )
+)
+```
+
+%% Output
+
+    Test Loss: 22.235297
+    
+    Test Accuracy of airplane: 52% (523/1000)
+    Test Accuracy of automobile: 84% (849/1000)
+    Test Accuracy of  bird: 34% (341/1000)
+    Test Accuracy of   cat: 43% (432/1000)
+    Test Accuracy of  deer: 66% (662/1000)
+    Test Accuracy of   dog: 44% (448/1000)
+    Test Accuracy of  frog: 74% (746/1000)
+    Test Accuracy of horse: 64% (647/1000)
+    Test Accuracy of  ship: 83% (836/1000)
+    Test Accuracy of truck: 64% (649/1000)
+    
+    Test Accuracy (Overall): 61% (6133/10000)
+
+%% Cell type:markdown id: tags:
+
+Then for the quantized model :
+
+%% Cell type:code id: tags:
+
+``` python
+# quantize model
+quantized_model = torch.quantization.quantize_dynamic(model, dtype=torch.qint8)
+
+# track test loss
+quantized_test_loss = 0.0
+quantized_class_correct = list(0.0 for i in range(10))
+quantized_class_total = list(0.0 for i in range(10))
+
+quantized_model.eval()
+# iterate over test data
+for data, target in test_loader:
+    # move tensors to GPU if CUDA is available
+    if train_on_gpu:
+        data, target = data.cuda(), target.cuda()
+    # forward pass: compute predicted outputs by passing inputs to the model
+    output = quantized_model(data)
+    # calculate the batch loss
+    loss = criterion(output, target)
+    # update test loss
+    quantized_test_loss += loss.item() * data.size(0)
+    # convert output probabilities to predicted class
+    _, pred = torch.max(output, 1)
+    # compare predictions to true label
+    correct_tensor = pred.eq(target.data.view_as(pred))
+    correct = (
+        np.squeeze(correct_tensor.numpy())
+        if not train_on_gpu
+        else np.squeeze(correct_tensor.cpu().numpy())
+    )
+    # calculate test accuracy for each object class
+    for i in range(batch_size):
+        label = target.data[i]
+        quantized_class_correct[label] += correct[i].item()
+        quantized_class_total[label] += 1
+
+# average test loss
+quantized_test_loss = quantized_test_loss / len(test_loader)
+print("Test Loss: {:.6f}\n".format(quantized_test_loss))
+
+for i in range(10):
+    if quantized_class_total[i] > 0:
+        print(
+            "Test Accuracy of %5s: %2d%% (%2d/%2d)"
+            % (
+                classes[i],
+                100 * quantized_class_correct[i] / quantized_class_total[i],
+                np.sum(quantized_class_correct[i]),
+                np.sum(quantized_class_total[i]),
+            )
+        )
+    else:
+        print("Test Accuracy of %5s: N/A (no training examples)" % (classes[i]))
+
+print(
+    "\nTest Accuracy (Overall): %2d%% (%2d/%2d)"
+    % (
+        100.0 * np.sum(quantized_class_correct) / np.sum(quantized_class_total),
+        np.sum(quantized_class_correct),
+        np.sum(quantized_class_total),
+    )
+)
+```
+
+%% Output
+
+    Test Loss: 22.242847
+    
+    Test Accuracy of airplane: 52% (523/1000)
+    Test Accuracy of automobile: 85% (853/1000)
+    Test Accuracy of  bird: 34% (342/1000)
+    Test Accuracy of   cat: 43% (430/1000)
+    Test Accuracy of  deer: 66% (660/1000)
+    Test Accuracy of   dog: 45% (452/1000)
+    Test Accuracy of  frog: 74% (749/1000)
+    Test Accuracy of horse: 64% (649/1000)
+    Test Accuracy of  ship: 83% (835/1000)
+    Test Accuracy of truck: 64% (645/1000)
+    
+    Test Accuracy (Overall): 61% (6138/10000)
+
+%% Cell type:markdown id: tags:
+
+The result is that the test accuracy is really similar for the initial model and for the quantized model.
+
 %% Cell type:markdown id:201470f9 tags:
 
 ## Exercise 3: working with pre-trained models.
 
 PyTorch offers several pre-trained models https://pytorch.org/vision/0.8/models.html
 We will use ResNet50 trained on ImageNet dataset (https://www.image-net.org/index.php). Use the following code with the files `imagenet-simple-labels.json` that contains the imagenet labels and the image dog.png that we will use as test.
 
 %% Cell type:code id:b4d13080 tags:
 
 ``` python
 import json
 from PIL import Image
 
 # Choose an image to pass through the model
 test_image = "dog.png"
 
 # Configure matplotlib for pretty inline plots
 #%matplotlib inline
 #%config InlineBackend.figure_format = 'retina'
 
 # Prepare the labels
 with open("imagenet-simple-labels.json") as f:
     labels = json.load(f)
 
 # First prepare the transformations: resize the image to what the model was trained on and convert it to a tensor
 data_transform = transforms.Compose(
     [
         transforms.Resize((224, 224)),
         transforms.ToTensor(),
         transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]),
     ]
 )
 # Load the image
 
 image = Image.open(test_image)
 plt.imshow(image), plt.xticks([]), plt.yticks([])
 
 # Now apply the transformation, expand the batch dimension, and send the image to the GPU
 # image = data_transform(image).unsqueeze(0).cuda()
 image = data_transform(image).unsqueeze(0)
 
 # Download the model if it's not there already. It will take a bit on the first run, after that it's fast
 model = models.resnet50(pretrained=True)
 # Send the model to the GPU
 # model.cuda()
 # Set layers such as dropout and batchnorm in evaluation mode
 model.eval()
 
 # Get the 1000-dimensional model output
 out = model(image)
 # Find the predicted class
 print("Predicted class is: {}".format(labels[out.argmax()]))
 ```
 
 %% Cell type:markdown id:184cfceb tags:
 
 Experiments:
 
 Study the code and the results obtained. Possibly add other images downloaded from the internet.
 
 What is the size of the model? Quantize it and then check if the model is still able to correctly classify the other images.
 
 Experiment with other pre-trained CNN models.
 
 
 
 %% Cell type:markdown id:5d57da4b tags:
 
 ## Exercise 4: Transfer Learning
 
 
 For this work, we will use a pre-trained model (ResNet18) as a descriptor extractor and will refine the classification by training only the last fully connected layer of the network. Thus, the output layer of the pre-trained network will be replaced by a layer adapted to the new classes to be recognized which will be in our case ants and bees.
 Download and unzip in your working directory the dataset available at the address :
 
 https://download.pytorch.org/tutorial/hymenoptera_data.zip
 
 Execute the following code in order to display some images of the dataset.
 
 %% Cell type:code id:be2d31f5 tags:
 
 ``` python
 import os
 
 import matplotlib.pyplot as plt
 import numpy as np
 import torch
 import torchvision
 from torchvision import datasets, transforms
 
 # Data augmentation and normalization for training
 # Just normalization for validation
 data_transforms = {
     "train": transforms.Compose(
         [
             transforms.RandomResizedCrop(
                 224
             ),  # ImageNet models were trained on 224x224 images
             transforms.RandomHorizontalFlip(),  # flip horizontally 50% of the time - increases train set variability
             transforms.ToTensor(),  # convert it to a PyTorch tensor
             transforms.Normalize(
                 [0.485, 0.456, 0.406], [0.229, 0.224, 0.225]
             ),  # ImageNet models expect this norm
         ]
     ),
     "val": transforms.Compose(
         [
             transforms.Resize(256),
             transforms.CenterCrop(224),
             transforms.ToTensor(),
             transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]),
         ]
     ),
 }
 
 data_dir = "hymenoptera_data"
 # Create train and validation datasets and loaders
 image_datasets = {
     x: datasets.ImageFolder(os.path.join(data_dir, x), data_transforms[x])
     for x in ["train", "val"]
 }
 dataloaders = {
     x: torch.utils.data.DataLoader(
         image_datasets[x], batch_size=4, shuffle=True, num_workers=0
     )
     for x in ["train", "val"]
 }
 dataset_sizes = {x: len(image_datasets[x]) for x in ["train", "val"]}
 class_names = image_datasets["train"].classes
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 
 # Helper function for displaying images
 def imshow(inp, title=None):
     """Imshow for Tensor."""
     inp = inp.numpy().transpose((1, 2, 0))
     mean = np.array([0.485, 0.456, 0.406])
     std = np.array([0.229, 0.224, 0.225])
 
     # Un-normalize the images
     inp = std * inp + mean
     # Clip just in case
     inp = np.clip(inp, 0, 1)
     plt.imshow(inp)
     if title is not None:
         plt.title(title)
     plt.pause(0.001)  # pause a bit so that plots are updated
     plt.show()
 
 
 # Get a batch of training data
 inputs, classes = next(iter(dataloaders["train"]))
 
 # Make a grid from batch
 out = torchvision.utils.make_grid(inputs)
 
 imshow(out, title=[class_names[x] for x in classes])
 
 ```
 
 %% Cell type:markdown id:bbd48800 tags:
 
 Now, execute the following code which uses a pre-trained model ResNet18 having replaced the output layer for the ants/bees classification and performs the model training by only changing the weights of this output layer.
 
 %% Cell type:code id:572d824c tags:
 
 ``` python
 import copy
 import os
 import time
 
 import matplotlib.pyplot as plt
 import numpy as np
 import torch
 import torch.nn as nn
 import torch.optim as optim
 import torchvision
 from torch.optim import lr_scheduler
 from torchvision import datasets, transforms
 
 # Data augmentation and normalization for training
 # Just normalization for validation
 data_transforms = {
     "train": transforms.Compose(
         [
             transforms.RandomResizedCrop(
                 224
             ),  # ImageNet models were trained on 224x224 images
             transforms.RandomHorizontalFlip(),  # flip horizontally 50% of the time - increases train set variability
             transforms.ToTensor(),  # convert it to a PyTorch tensor
             transforms.Normalize(
                 [0.485, 0.456, 0.406], [0.229, 0.224, 0.225]
             ),  # ImageNet models expect this norm
         ]
     ),
     "val": transforms.Compose(
         [
             transforms.Resize(256),
             transforms.CenterCrop(224),
             transforms.ToTensor(),
             transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]),
         ]
     ),
 }
 
 data_dir = "hymenoptera_data"
 # Create train and validation datasets and loaders
 image_datasets = {
     x: datasets.ImageFolder(os.path.join(data_dir, x), data_transforms[x])
     for x in ["train", "val"]
 }
 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"]}
 class_names = image_datasets["train"].classes
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 
 # Helper function for displaying images
 def imshow(inp, title=None):
     """Imshow for Tensor."""
     inp = inp.numpy().transpose((1, 2, 0))
     mean = np.array([0.485, 0.456, 0.406])
     std = np.array([0.229, 0.224, 0.225])
 
     # Un-normalize the images
     inp = std * inp + mean
     # Clip just in case
     inp = np.clip(inp, 0, 1)
     plt.imshow(inp)
     if title is not None:
         plt.title(title)
     plt.pause(0.001)  # pause a bit so that plots are updated
     plt.show()
 
 
 # Get a batch of training data
 # inputs, classes = next(iter(dataloaders['train']))
 
 # Make a grid from batch
 # out = torchvision.utils.make_grid(inputs)
 
 # imshow(out, title=[class_names[x] for x in classes])
 # training
 
 
 def train_model(model, criterion, optimizer, scheduler, num_epochs=25):
     since = time.time()
 
     best_model_wts = copy.deepcopy(model.state_dict())
     best_acc = 0.0
 
     epoch_time = []  # we'll keep track of the time needed for each epoch
 
     for epoch in range(num_epochs):
         epoch_start = time.time()
         print("Epoch {}/{}".format(epoch + 1, num_epochs))
         print("-" * 10)
 
         # Each epoch has a training and validation phase
         for phase in ["train", "val"]:
             if phase == "train":
                 scheduler.step()
                 model.train()  # Set model to training mode
             else:
                 model.eval()  # Set model to evaluate mode
 
             running_loss = 0.0
             running_corrects = 0
 
             # Iterate over data.
             for inputs, labels in dataloaders[phase]:
                 inputs = inputs.to(device)
                 labels = labels.to(device)
 
                 # zero the parameter gradients
                 optimizer.zero_grad()
 
                 # Forward
                 # Track history if only in training phase
                 with torch.set_grad_enabled(phase == "train"):
                     outputs = model(inputs)
                     _, preds = torch.max(outputs, 1)
                     loss = criterion(outputs, labels)
 
                     # backward + optimize only if in training phase
                     if phase == "train":
                         loss.backward()
                         optimizer.step()
 
                 # Statistics
                 running_loss += loss.item() * inputs.size(0)
                 running_corrects += torch.sum(preds == labels.data)
 
             epoch_loss = running_loss / dataset_sizes[phase]
             epoch_acc = running_corrects.double() / dataset_sizes[phase]
 
             print("{} Loss: {:.4f} Acc: {:.4f}".format(phase, epoch_loss, epoch_acc))
 
             # Deep copy the model
             if phase == "val" and epoch_acc > best_acc:
                 best_acc = epoch_acc
                 best_model_wts = copy.deepcopy(model.state_dict())
 
         # Add the epoch time
         t_epoch = time.time() - epoch_start
         epoch_time.append(t_epoch)
         print()
 
     time_elapsed = time.time() - since
     print(
         "Training complete in {:.0f}m {:.0f}s".format(
             time_elapsed // 60, time_elapsed % 60
         )
     )
     print("Best val Acc: {:4f}".format(best_acc))
 
     # Load best model weights
     model.load_state_dict(best_model_wts)
     return model, epoch_time
 
 
 # Download a pre-trained ResNet18 model and freeze its weights
 model = torchvision.models.resnet18(pretrained=True)
 for param in model.parameters():
     param.requires_grad = False
 
 # Replace the final fully connected layer
 # Parameters of newly constructed modules have requires_grad=True by default
 num_ftrs = model.fc.in_features
 model.fc = nn.Linear(num_ftrs, 2)
 # Send the model to the GPU
 model = model.to(device)
 # Set the loss function
 criterion = nn.CrossEntropyLoss()
 
 # Observe that only the parameters of the final layer are being optimized
 optimizer_conv = optim.SGD(model.fc.parameters(), lr=0.001, momentum=0.9)
 exp_lr_scheduler = lr_scheduler.StepLR(optimizer_conv, step_size=7, gamma=0.1)
 model, epoch_time = train_model(
     model, criterion, optimizer_conv, exp_lr_scheduler, num_epochs=10
 )
 ```
 
 %% Cell type:markdown id:bbd48800 tags:
 
 Experiments:
 Study the code and the results obtained.
 
 Modify the code and add an "eval_model" function to allow
 the evaluation of the model on a test set (different from the learning and validation sets used during the learning phase). Study the results obtained.
 
 Now modify the code to replace the current classification layer with a set of two layers using a "relu" activation function for the middle layer, and the "dropout" mechanism for both layers. Renew the experiments and study the results obtained.
 
 Apply ther quantization (post and quantization aware) and evaluate impact on model size and accuracy.
 
 %% Cell type:markdown id:04a263f0 tags:
 
 ## Optional
 
 Try this at home!!
 
 
 Pytorch offers a framework to export a given CNN to your selfphone (either android or iOS). Have a look at the tutorial https://pytorch.org/mobile/home/
 
 The Exercise consists in deploying the CNN of Exercise 4 in your phone and then test it on live.
 
 
 %% Cell type:markdown id:fe954ce4 tags:
 
 ## Author
 
 Alberto BOSIO - Ph. D.