Compare revisions

Vincent Simon · 603ab641 · 603ab641 · 603ab641
--- a/TD2 Deep Learning.ipynb
+++ b/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 Wednesday, December 4, 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
 ```

 %% 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([[ 0.1159, -0.1632,  0.6574,  1.5902, -0.4352, -1.1418,  0.8810,  0.0847,
+             -2.4290, -0.0911],
+            [ 0.3400, -1.9606, -0.0214, -0.1179, -0.3917, -0.3592,  0.5251,  0.4169,
+             -0.8085, -0.2057],
+            [-0.7488,  0.7606,  0.1129, -2.6223, -0.5739, -0.4979,  2.0387,  0.1628,
+              1.1597, -0.9275],
+            [-1.5324,  1.4420,  0.9108,  0.4737,  0.3852, -1.1974,  1.7244,  1.3268,
+              1.4552,  0.5241],
+            [-0.3818, -0.4960, -1.5574, -0.8755,  1.2589,  0.8939,  0.0385, -2.5047,
+              0.6804, -0.1951],
+            [ 0.1988,  0.9232, -1.3031,  1.8143,  0.0756,  1.2082, -1.1921,  0.0647,
+              0.1529,  0.4644],
+            [ 1.8262,  0.6831, -0.1683, -0.8331, -0.5271, -0.2069,  0.5703,  1.7226,
+             -0.6655, -0.4297],
+            [-0.0630, -0.2216,  2.2132, -0.8788,  2.8345, -0.0534, -1.7918, -0.6061,
+             -0.2461,  0.4126],
+            [ 0.3832, -0.2473, -1.1898,  2.3250,  0.1655, -0.4416, -0.4937, -0.1714,
+              0.6682, -0.7186],
+            [-0.5843,  1.7539,  0.4247,  0.5102, -1.2161,  0.2732,  1.8955,  1.5722,
+              0.9527,  0.2717],
+            [-1.2976,  0.2779, -0.8085,  0.0037, -1.4008, -1.3840,  0.1210,  0.5056,
+              0.6006, -1.5492],
+            [-0.1415, -0.8489,  0.3045,  2.3843,  1.4306, -0.5467, -0.2279,  0.2920,
+              1.5270, -1.5247],
+            [-0.8661, -0.3661,  0.3478, -0.5955,  1.0730, -2.1341, -0.8818,  0.2842,
+              0.8046,  0.4630],
+            [-0.1986,  1.3981, -0.3965, -0.6231,  2.5136,  0.1703, -1.0520, -0.4539,
+             -1.8835, -0.1314]])
+    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
+
+    Downloading https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz to data\cifar-10-python.tar.gz
+
+    100.0%
+
+    Extracting data\cifar-10-python.tar.gz to data
+    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:02ad19e0 tags:
+
+Creating the model to answer question 1:
+
+%% Cell type:code id:9f3145ca tags:
+
+``` python
+
+
+class Net_3Conv_3lin(nn.Module):
+    def __init__(self):
+        super(Net_3Conv_3lin, self).__init__()
+        self.conv1 = nn.Conv2d(kernel_size=3 ,padding=1 ,in_channels=3, out_channels =16)
+        self.conv2 = nn.Conv2d(kernel_size=3 ,padding=1,in_channels=16, out_channels =32 )
+        self.conv3 = nn.Conv2d(kernel_size=3 ,padding=1,in_channels=32, out_channels =64 )
+        self.pool = nn.MaxPool2d(2, 2)
+        self.fc1 = nn.Linear(1024,512)
+        self.fc2 = nn.Linear(512,64)
+        self.fc3 = nn.Linear(64,10)
+        self.p = 0.3
+        self.dropout = nn.Dropout(self.p)
+
+    def forward(self, x):
+        x = self.pool(F.relu(self.conv1(x)))
+        x = self.pool(F.relu(self.conv2(x)))
+        x = self.pool(F.relu(self.conv3(x)))
+        x = x.view(-1, 64 * 4 * 4)
+        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 = Net_3Conv_3lin()
+print(model)
+# move tensors to GPU if CUDA is available
+if train_on_gpu:
+    model.cuda()
+```
+
+%% Output
+
+    Net_3Conv_3lin(
+      (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=0.3, inplace=False)
+    )
+
+%% Cell type:markdown id: tags:
+
+Creating a model for training aware quatization
+
+%% Cell type:code id:71bb37b5 tags:
+
+``` python
+
+
+class Net_3Conv_3lin_Quant(nn.Module):
+    def __init__(self):
+        super(Net_3Conv_3lin_Quant, self).__init__()
+        self.quant = torch.ao.quantization.QuantStub()
+        self.conv1 = nn.Conv2d(kernel_size=3 ,padding=1 ,in_channels=3, out_channels =16)
+        self.conv2 = nn.Conv2d(kernel_size=3 ,padding=1,in_channels=16, out_channels =32 )
+        self.conv3 = nn.Conv2d(kernel_size=3 ,padding=1,in_channels=32, out_channels =64 )
+        self.pool = nn.MaxPool2d(2, 2)
+        self.fc1 = nn.Linear(1024,512)
+        self.fc2 = nn.Linear(512,64)
+        self.fc3 = nn.Linear(64,10)
+        self.p = 0.3
+        self.dropout = nn.Dropout(self.p)
+
+        self.dequant = torch.ao.quantization.DeQuantStub()
+
+    def forward(self, x):
+        x = self.quant(x)
+        x = self.pool(F.relu(self.conv1(x)))
+        x = self.pool(F.relu(self.conv2(x)))
+        x = self.pool(F.relu(self.conv3(x)))
+        x = x.view(-1, 64 * 4 * 4)
+        x = self.dropout(F.relu(self.fc1(x)))
+        x = self.dropout(F.relu(self.fc2(x)))
+        x = self.fc3(x)
+        x = self.dequant(x)
+        return x
+
+
+# create a complete CNN
+model = Net_3Conv_3lin_Quant()
+print(model)
+# move tensors to GPU if CUDA is available
+if train_on_gpu:
+    model.cuda()
+```
+
 %% 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
+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: 45.872737 	Validation Loss: 44.489702
+    Validation loss decreased (inf --> 44.489702).  Saving model ...
+    Epoch: 1 	Training Loss: 40.061390 	Validation Loss: 36.100894
+    Validation loss decreased (44.489702 --> 36.100894).  Saving model ...
+    Epoch: 2 	Training Loss: 34.138645 	Validation Loss: 31.113222
+    Validation loss decreased (36.100894 --> 31.113222).  Saving model ...
+    Epoch: 3 	Training Loss: 30.507358 	Validation Loss: 28.677294
+    Validation loss decreased (31.113222 --> 28.677294).  Saving model ...
+    Epoch: 4 	Training Loss: 28.486533 	Validation Loss: 26.945912
+    Validation loss decreased (28.677294 --> 26.945912).  Saving model ...
+    Epoch: 5 	Training Loss: 26.688108 	Validation Loss: 25.087968
+    Validation loss decreased (26.945912 --> 25.087968).  Saving model ...
+    Epoch: 6 	Training Loss: 24.921677 	Validation Loss: 23.867951
+    Validation loss decreased (25.087968 --> 23.867951).  Saving model ...
+    Epoch: 7 	Training Loss: 23.352516 	Validation Loss: 21.780911
+    Validation loss decreased (23.867951 --> 21.780911).  Saving model ...
+    Epoch: 8 	Training Loss: 21.794870 	Validation Loss: 21.096160
+    Validation loss decreased (21.780911 --> 21.096160).  Saving model ...
+    Epoch: 9 	Training Loss: 20.547996 	Validation Loss: 19.826372
+    Validation loss decreased (21.096160 --> 19.826372).  Saving model ...
+    Epoch: 10 	Training Loss: 19.401682 	Validation Loss: 19.596204
+    Validation loss decreased (19.826372 --> 19.596204).  Saving model ...
+    Epoch: 11 	Training Loss: 18.340276 	Validation Loss: 18.632437
+    Validation loss decreased (19.596204 --> 18.632437).  Saving model ...
+    Epoch: 12 	Training Loss: 17.266555 	Validation Loss: 17.758480
+    Validation loss decreased (18.632437 --> 17.758480).  Saving model ...
+    Epoch: 13 	Training Loss: 16.353216 	Validation Loss: 17.932480
+    Epoch: 14 	Training Loss: 15.507940 	Validation Loss: 16.795444
+    Validation loss decreased (17.758480 --> 16.795444).  Saving model ...
+    Epoch: 15 	Training Loss: 14.657860 	Validation Loss: 16.382975
+    Validation loss decreased (16.795444 --> 16.382975).  Saving model ...
+    Epoch: 16 	Training Loss: 13.861092 	Validation Loss: 16.670121
+    Epoch: 17 	Training Loss: 12.984836 	Validation Loss: 16.962824
+    Epoch: 18 	Training Loss: 12.257837 	Validation Loss: 16.400703
+    Epoch: 19 	Training Loss: 11.436899 	Validation Loss: 16.729391
+    Epoch: 20 	Training Loss: 10.915463 	Validation Loss: 16.299635
+    Validation loss decreased (16.382975 --> 16.299635).  Saving model ...
+    Epoch: 21 	Training Loss: 10.233074 	Validation Loss: 16.345074
+    Epoch: 22 	Training Loss: 9.549847 	Validation Loss: 16.705205
+    Epoch: 23 	Training Loss: 8.865565 	Validation Loss: 16.583533
+    Epoch: 24 	Training Loss: 8.210216 	Validation Loss: 17.112398
+    Epoch: 25 	Training Loss: 7.781697 	Validation Loss: 17.192360
+
+    ---------------------------------------------------------------------------
+    KeyboardInterrupt                         Traceback (most recent call last)
+Cell     In[34], line 41
+         39     data, target = data.cuda(), target.cuda()
+         40 # Forward pass: compute predicted outputs by passing inputs to the model
+    ---> 41 output = model(data)
+         42 # Calculate the batch loss
+         43 loss = criterion(output, target)
+File     c:\Users\xxpod\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\nn\modules\module.py:1736, in Module._wrapped_call_impl(self, *args, **kwargs)
+       1734     return self._compiled_call_impl(*args, **kwargs)  # type: ignore[misc]
+       1735 else:
+    -> 1736     return self._call_impl(*args, **kwargs)
+File     c:\Users\xxpod\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\nn\modules\module.py:1747, in Module._call_impl(self, *args, **kwargs)
+       1742 # If we don't have any hooks, we want to skip the rest of the logic in
+       1743 # this function, and just call forward.
+       1744 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
+       1745         or _global_backward_pre_hooks or _global_backward_hooks
+       1746         or _global_forward_hooks or _global_forward_pre_hooks):
+    -> 1747     return forward_call(*args, **kwargs)
+       1749 result = None
+       1750 called_always_called_hooks = set()
+Cell     In[33], line 16, in Net_3Conv_3lin.forward(self, x)
+         14 def forward(self, x):
+         15     x = self.pool(F.relu(self.conv1(x)))
+    ---> 16     x = self.pool(F.relu(self.conv2(x)))
+         17     x = self.pool(F.relu(self.conv3(x)))
+         18     x = x.view(-1, 64 * 4 * 4)
+File     c:\Users\xxpod\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\nn\modules\module.py:1736, in Module._wrapped_call_impl(self, *args, **kwargs)
+       1734     return self._compiled_call_impl(*args, **kwargs)  # type: ignore[misc]
+       1735 else:
+    -> 1736     return self._call_impl(*args, **kwargs)
+File     c:\Users\xxpod\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\nn\modules\module.py:1747, in Module._call_impl(self, *args, **kwargs)
+       1742 # If we don't have any hooks, we want to skip the rest of the logic in
+       1743 # this function, and just call forward.
+       1744 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
+       1745         or _global_backward_pre_hooks or _global_backward_hooks
+       1746         or _global_forward_hooks or _global_forward_pre_hooks):
+    -> 1747     return forward_call(*args, **kwargs)
+       1749 result = None
+       1750 called_always_called_hooks = set()
+File     c:\Users\xxpod\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\nn\modules\pooling.py:213, in MaxPool2d.forward(self, input)
+        212 def forward(self, input: Tensor):
+    --> 213     return F.max_pool2d(
+        214         input,
+        215         self.kernel_size,
+        216         self.stride,
+        217         self.padding,
+        218         self.dilation,
+        219         ceil_mode=self.ceil_mode,
+        220         return_indices=self.return_indices,
+        221     )
+File     c:\Users\xxpod\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\_jit_internal.py:624, in boolean_dispatch.<locals>.fn(*args, **kwargs)
+        622     return if_true(*args, **kwargs)
+        623 else:
+    --> 624     return if_false(*args, **kwargs)
+File     c:\Users\xxpod\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\nn\functional.py:830, in _max_pool2d(input, kernel_size, stride, padding, dilation, ceil_mode, return_indices)
+        828 if stride is None:
+        829     stride = torch.jit.annotate(List[int], [])
+    --> 830 return torch.max_pool2d(input, kernel_size, stride, padding, dilation, ceil_mode)
+    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)
+print(model)
+plt.plot(range(len(train_loss_list)), train_loss_list)
 plt.xlabel("Epoch")
 plt.ylabel("Loss")
-plt.title("Performance of Model 1")
+plt.title("Performance of Model 2")
 plt.show()
 ```

+%% Output
+
+    Net_3Conv_3lin(
+      (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=0.3, inplace=False)
+    )
+
+
+
 %% 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
+
+    C:\Users\xxpod\AppData\Local\Temp\ipykernel_18828\3291884398.py:1: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
+      model.load_state_dict(torch.load("./model_cifar.pt"))
+
+    Test Loss: 16.123924
+    
+    Test Accuracy of airplane: 81% (810/1000)
+    Test Accuracy of automobile: 85% (855/1000)
+    Test Accuracy of  bird: 63% (633/1000)
+    Test Accuracy of   cat: 52% (525/1000)
+    Test Accuracy of  deer: 69% (695/1000)
+    Test Accuracy of   dog: 71% (717/1000)
+    Test Accuracy of  frog: 77% (772/1000)
+    Test Accuracy of horse: 77% (772/1000)
+    Test Accuracy of  ship: 84% (843/1000)
+    Test Accuracy of truck: 76% (765/1000)
+    
+    Test Accuracy (Overall): 73% (7387/10000)
+
 %% 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.

+ANSWER: The model is built above and named Net_Conv3_Lin3
+
+
+Results for the previous model :
+
+we osberve overfitting from about the 10nth Epoch - validation loss plateaued at 22 but training loss kept on decreasing to 10 , as can be seen from the training logs:
+
+Epoch: 7 	Training Loss: 23.183946 	Validation Loss: 24.331222
+Validation loss decreased (25.691083 --> 24.331222).  Saving model ...
+
+Epoch: 8 	Training Loss: 22.215979 	Validation Loss: 23.632853
+Validation loss decreased (24.331222 --> 23.632853).  Saving model ...
+
+Epoch: 9 	Training Loss: 21.408623 	Validation Loss: 23.475442
+Validation loss decreased (23.632853 --> 23.475442).  Saving model ...
+
+Epoch: 10 	Training Loss: 20.637072 	Validation Loss: 23.639358
+
+Epoch: 11 	Training Loss: 19.877338 	Validation Loss: 22.408472
+Validation loss decreased (23.475442 --> 22.408472).  Saving model ...
+
+Epoch: 12 	Training Loss: 19.188079 	Validation Loss: 23.296445
+
+Epoch: 13 	Training Loss: 18.647543 	Validation Loss: 22.897815
+
+Epoch: 14 	Training Loss: 17.989626 	Validation Loss: 22.755968
+
+the performance is as follow: ![alt text](perf_model_1.png)
+
+and the final accuries were:
+![Accuracies](final_accuracy_first_model.png)
+
+
+SECOND MODEL:
+
+for the second model, the validation loss goes lower, thougth in addition to the architectural changes, there are also just more weigth and it is longer to train.
+
+we archieve a valisation loss of 16, and the model is still improving after a larger number of epoch ( 20 vs 10)
+
+
+here are the final accuracies:
+Test Loss: 16.123924
+
+Test Accuracy of airplane: 81% (810/1000)
+Test Accuracy of automobile: 85% (855/1000)
+Test Accuracy of  bird: 63% (633/1000)
+Test Accuracy of   cat: 52% (525/1000)
+Test Accuracy of  deer: 69% (695/1000)
+Test Accuracy of   dog: 71% (717/1000)
+Test Accuracy of  frog: 77% (772/1000)
+Test Accuracy of horse: 77% (772/1000)
+Test Accuracy of  ship: 84% (843/1000)
+Test Accuracy of truck: 76% (765/1000)
+
+Test Accuracy (Overall): 73% (7387/10000)
+
 %% 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")
 ```

 %% 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")
 ```

 %% 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:code id:1458a562 tags:
+
+``` python
+model = Net_3Conv_3lin_Quant()
+print(model)
+
+
+#prepare for fusion
+model.eval()
+#model.qconfig = torch.ao.quantization.get_default_qat_config('86')
+#model_fused = torch.ao.quantization.fuse_modules(model,[['conv', 'relu',]])
+
+model.train()
+model_prepared = torch.ao.quantization.prepare_qat(model)
+training_loop(model_prepared)
+model_prepared.eval()
+model_quantized = torch.ao.quantization.convert(model_prepared)
+
+evaluate(model_quantized)
+print_size_of_model(model_quantized)
+
+```
+
 %% 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: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.

 %% 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 Wednesday, December 4, 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
 ```

 %% 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([[ 0.1159, -0.1632,  0.6574,  1.5902, -0.4352, -1.1418,  0.8810,  0.0847,
+             -2.4290, -0.0911],
+            [ 0.3400, -1.9606, -0.0214, -0.1179, -0.3917, -0.3592,  0.5251,  0.4169,
+             -0.8085, -0.2057],
+            [-0.7488,  0.7606,  0.1129, -2.6223, -0.5739, -0.4979,  2.0387,  0.1628,
+              1.1597, -0.9275],
+            [-1.5324,  1.4420,  0.9108,  0.4737,  0.3852, -1.1974,  1.7244,  1.3268,
+              1.4552,  0.5241],
+            [-0.3818, -0.4960, -1.5574, -0.8755,  1.2589,  0.8939,  0.0385, -2.5047,
+              0.6804, -0.1951],
+            [ 0.1988,  0.9232, -1.3031,  1.8143,  0.0756,  1.2082, -1.1921,  0.0647,
+              0.1529,  0.4644],
+            [ 1.8262,  0.6831, -0.1683, -0.8331, -0.5271, -0.2069,  0.5703,  1.7226,
+             -0.6655, -0.4297],
+            [-0.0630, -0.2216,  2.2132, -0.8788,  2.8345, -0.0534, -1.7918, -0.6061,
+             -0.2461,  0.4126],
+            [ 0.3832, -0.2473, -1.1898,  2.3250,  0.1655, -0.4416, -0.4937, -0.1714,
+              0.6682, -0.7186],
+            [-0.5843,  1.7539,  0.4247,  0.5102, -1.2161,  0.2732,  1.8955,  1.5722,
+              0.9527,  0.2717],
+            [-1.2976,  0.2779, -0.8085,  0.0037, -1.4008, -1.3840,  0.1210,  0.5056,
+              0.6006, -1.5492],
+            [-0.1415, -0.8489,  0.3045,  2.3843,  1.4306, -0.5467, -0.2279,  0.2920,
+              1.5270, -1.5247],
+            [-0.8661, -0.3661,  0.3478, -0.5955,  1.0730, -2.1341, -0.8818,  0.2842,
+              0.8046,  0.4630],
+            [-0.1986,  1.3981, -0.3965, -0.6231,  2.5136,  0.1703, -1.0520, -0.4539,
+             -1.8835, -0.1314]])
+    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
+
+    Downloading https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz to data\cifar-10-python.tar.gz
+
+    100.0%
+
+    Extracting data\cifar-10-python.tar.gz to data
+    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:02ad19e0 tags:
+
+Creating the model to answer question 1:
+
+%% Cell type:code id:9f3145ca tags:
+
+``` python
+
+
+class Net_3Conv_3lin(nn.Module):
+    def __init__(self):
+        super(Net_3Conv_3lin, self).__init__()
+        self.conv1 = nn.Conv2d(kernel_size=3 ,padding=1 ,in_channels=3, out_channels =16)
+        self.conv2 = nn.Conv2d(kernel_size=3 ,padding=1,in_channels=16, out_channels =32 )
+        self.conv3 = nn.Conv2d(kernel_size=3 ,padding=1,in_channels=32, out_channels =64 )
+        self.pool = nn.MaxPool2d(2, 2)
+        self.fc1 = nn.Linear(1024,512)
+        self.fc2 = nn.Linear(512,64)
+        self.fc3 = nn.Linear(64,10)
+        self.p = 0.3
+        self.dropout = nn.Dropout(self.p)
+
+    def forward(self, x):
+        x = self.pool(F.relu(self.conv1(x)))
+        x = self.pool(F.relu(self.conv2(x)))
+        x = self.pool(F.relu(self.conv3(x)))
+        x = x.view(-1, 64 * 4 * 4)
+        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 = Net_3Conv_3lin()
+print(model)
+# move tensors to GPU if CUDA is available
+if train_on_gpu:
+    model.cuda()
+```
+
+%% Output
+
+    Net_3Conv_3lin(
+      (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=0.3, inplace=False)
+    )
+
+%% Cell type:markdown id: tags:
+
+Creating a model for training aware quatization
+
+%% Cell type:code id:71bb37b5 tags:
+
+``` python
+
+
+class Net_3Conv_3lin_Quant(nn.Module):
+    def __init__(self):
+        super(Net_3Conv_3lin_Quant, self).__init__()
+        self.quant = torch.ao.quantization.QuantStub()
+        self.conv1 = nn.Conv2d(kernel_size=3 ,padding=1 ,in_channels=3, out_channels =16)
+        self.conv2 = nn.Conv2d(kernel_size=3 ,padding=1,in_channels=16, out_channels =32 )
+        self.conv3 = nn.Conv2d(kernel_size=3 ,padding=1,in_channels=32, out_channels =64 )
+        self.pool = nn.MaxPool2d(2, 2)
+        self.fc1 = nn.Linear(1024,512)
+        self.fc2 = nn.Linear(512,64)
+        self.fc3 = nn.Linear(64,10)
+        self.p = 0.3
+        self.dropout = nn.Dropout(self.p)
+
+        self.dequant = torch.ao.quantization.DeQuantStub()
+
+    def forward(self, x):
+        x = self.quant(x)
+        x = self.pool(F.relu(self.conv1(x)))
+        x = self.pool(F.relu(self.conv2(x)))
+        x = self.pool(F.relu(self.conv3(x)))
+        x = x.view(-1, 64 * 4 * 4)
+        x = self.dropout(F.relu(self.fc1(x)))
+        x = self.dropout(F.relu(self.fc2(x)))
+        x = self.fc3(x)
+        x = self.dequant(x)
+        return x
+
+
+# create a complete CNN
+model = Net_3Conv_3lin_Quant()
+print(model)
+# move tensors to GPU if CUDA is available
+if train_on_gpu:
+    model.cuda()
+```
+
 %% 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
+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: 45.872737 	Validation Loss: 44.489702
+    Validation loss decreased (inf --> 44.489702).  Saving model ...
+    Epoch: 1 	Training Loss: 40.061390 	Validation Loss: 36.100894
+    Validation loss decreased (44.489702 --> 36.100894).  Saving model ...
+    Epoch: 2 	Training Loss: 34.138645 	Validation Loss: 31.113222
+    Validation loss decreased (36.100894 --> 31.113222).  Saving model ...
+    Epoch: 3 	Training Loss: 30.507358 	Validation Loss: 28.677294
+    Validation loss decreased (31.113222 --> 28.677294).  Saving model ...
+    Epoch: 4 	Training Loss: 28.486533 	Validation Loss: 26.945912
+    Validation loss decreased (28.677294 --> 26.945912).  Saving model ...
+    Epoch: 5 	Training Loss: 26.688108 	Validation Loss: 25.087968
+    Validation loss decreased (26.945912 --> 25.087968).  Saving model ...
+    Epoch: 6 	Training Loss: 24.921677 	Validation Loss: 23.867951
+    Validation loss decreased (25.087968 --> 23.867951).  Saving model ...
+    Epoch: 7 	Training Loss: 23.352516 	Validation Loss: 21.780911
+    Validation loss decreased (23.867951 --> 21.780911).  Saving model ...
+    Epoch: 8 	Training Loss: 21.794870 	Validation Loss: 21.096160
+    Validation loss decreased (21.780911 --> 21.096160).  Saving model ...
+    Epoch: 9 	Training Loss: 20.547996 	Validation Loss: 19.826372
+    Validation loss decreased (21.096160 --> 19.826372).  Saving model ...
+    Epoch: 10 	Training Loss: 19.401682 	Validation Loss: 19.596204
+    Validation loss decreased (19.826372 --> 19.596204).  Saving model ...
+    Epoch: 11 	Training Loss: 18.340276 	Validation Loss: 18.632437
+    Validation loss decreased (19.596204 --> 18.632437).  Saving model ...
+    Epoch: 12 	Training Loss: 17.266555 	Validation Loss: 17.758480
+    Validation loss decreased (18.632437 --> 17.758480).  Saving model ...
+    Epoch: 13 	Training Loss: 16.353216 	Validation Loss: 17.932480
+    Epoch: 14 	Training Loss: 15.507940 	Validation Loss: 16.795444
+    Validation loss decreased (17.758480 --> 16.795444).  Saving model ...
+    Epoch: 15 	Training Loss: 14.657860 	Validation Loss: 16.382975
+    Validation loss decreased (16.795444 --> 16.382975).  Saving model ...
+    Epoch: 16 	Training Loss: 13.861092 	Validation Loss: 16.670121
+    Epoch: 17 	Training Loss: 12.984836 	Validation Loss: 16.962824
+    Epoch: 18 	Training Loss: 12.257837 	Validation Loss: 16.400703
+    Epoch: 19 	Training Loss: 11.436899 	Validation Loss: 16.729391
+    Epoch: 20 	Training Loss: 10.915463 	Validation Loss: 16.299635
+    Validation loss decreased (16.382975 --> 16.299635).  Saving model ...
+    Epoch: 21 	Training Loss: 10.233074 	Validation Loss: 16.345074
+    Epoch: 22 	Training Loss: 9.549847 	Validation Loss: 16.705205
+    Epoch: 23 	Training Loss: 8.865565 	Validation Loss: 16.583533
+    Epoch: 24 	Training Loss: 8.210216 	Validation Loss: 17.112398
+    Epoch: 25 	Training Loss: 7.781697 	Validation Loss: 17.192360
+
+    ---------------------------------------------------------------------------
+    KeyboardInterrupt                         Traceback (most recent call last)
+Cell     In[34], line 41
+         39     data, target = data.cuda(), target.cuda()
+         40 # Forward pass: compute predicted outputs by passing inputs to the model
+    ---> 41 output = model(data)
+         42 # Calculate the batch loss
+         43 loss = criterion(output, target)
+File     c:\Users\xxpod\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\nn\modules\module.py:1736, in Module._wrapped_call_impl(self, *args, **kwargs)
+       1734     return self._compiled_call_impl(*args, **kwargs)  # type: ignore[misc]
+       1735 else:
+    -> 1736     return self._call_impl(*args, **kwargs)
+File     c:\Users\xxpod\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\nn\modules\module.py:1747, in Module._call_impl(self, *args, **kwargs)
+       1742 # If we don't have any hooks, we want to skip the rest of the logic in
+       1743 # this function, and just call forward.
+       1744 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
+       1745         or _global_backward_pre_hooks or _global_backward_hooks
+       1746         or _global_forward_hooks or _global_forward_pre_hooks):
+    -> 1747     return forward_call(*args, **kwargs)
+       1749 result = None
+       1750 called_always_called_hooks = set()
+Cell     In[33], line 16, in Net_3Conv_3lin.forward(self, x)
+         14 def forward(self, x):
+         15     x = self.pool(F.relu(self.conv1(x)))
+    ---> 16     x = self.pool(F.relu(self.conv2(x)))
+         17     x = self.pool(F.relu(self.conv3(x)))
+         18     x = x.view(-1, 64 * 4 * 4)
+File     c:\Users\xxpod\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\nn\modules\module.py:1736, in Module._wrapped_call_impl(self, *args, **kwargs)
+       1734     return self._compiled_call_impl(*args, **kwargs)  # type: ignore[misc]
+       1735 else:
+    -> 1736     return self._call_impl(*args, **kwargs)
+File     c:\Users\xxpod\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\nn\modules\module.py:1747, in Module._call_impl(self, *args, **kwargs)
+       1742 # If we don't have any hooks, we want to skip the rest of the logic in
+       1743 # this function, and just call forward.
+       1744 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
+       1745         or _global_backward_pre_hooks or _global_backward_hooks
+       1746         or _global_forward_hooks or _global_forward_pre_hooks):
+    -> 1747     return forward_call(*args, **kwargs)
+       1749 result = None
+       1750 called_always_called_hooks = set()
+File     c:\Users\xxpod\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\nn\modules\pooling.py:213, in MaxPool2d.forward(self, input)
+        212 def forward(self, input: Tensor):
+    --> 213     return F.max_pool2d(
+        214         input,
+        215         self.kernel_size,
+        216         self.stride,
+        217         self.padding,
+        218         self.dilation,
+        219         ceil_mode=self.ceil_mode,
+        220         return_indices=self.return_indices,
+        221     )
+File     c:\Users\xxpod\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\_jit_internal.py:624, in boolean_dispatch.<locals>.fn(*args, **kwargs)
+        622     return if_true(*args, **kwargs)
+        623 else:
+    --> 624     return if_false(*args, **kwargs)
+File     c:\Users\xxpod\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\nn\functional.py:830, in _max_pool2d(input, kernel_size, stride, padding, dilation, ceil_mode, return_indices)
+        828 if stride is None:
+        829     stride = torch.jit.annotate(List[int], [])
+    --> 830 return torch.max_pool2d(input, kernel_size, stride, padding, dilation, ceil_mode)
+    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)
+print(model)
+plt.plot(range(len(train_loss_list)), train_loss_list)
 plt.xlabel("Epoch")
 plt.ylabel("Loss")
-plt.title("Performance of Model 1")
+plt.title("Performance of Model 2")
 plt.show()
 ```

+%% Output
+
+    Net_3Conv_3lin(
+      (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=0.3, inplace=False)
+    )
+
+
+
 %% 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
+
+    C:\Users\xxpod\AppData\Local\Temp\ipykernel_18828\3291884398.py:1: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
+      model.load_state_dict(torch.load("./model_cifar.pt"))
+
+    Test Loss: 16.123924
+    
+    Test Accuracy of airplane: 81% (810/1000)
+    Test Accuracy of automobile: 85% (855/1000)
+    Test Accuracy of  bird: 63% (633/1000)
+    Test Accuracy of   cat: 52% (525/1000)
+    Test Accuracy of  deer: 69% (695/1000)
+    Test Accuracy of   dog: 71% (717/1000)
+    Test Accuracy of  frog: 77% (772/1000)
+    Test Accuracy of horse: 77% (772/1000)
+    Test Accuracy of  ship: 84% (843/1000)
+    Test Accuracy of truck: 76% (765/1000)
+    
+    Test Accuracy (Overall): 73% (7387/10000)
+
 %% 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.

+ANSWER: The model is built above and named Net_Conv3_Lin3
+
+
+Results for the previous model :
+
+we osberve overfitting from about the 10nth Epoch - validation loss plateaued at 22 but training loss kept on decreasing to 10 , as can be seen from the training logs:
+
+Epoch: 7 	Training Loss: 23.183946 	Validation Loss: 24.331222
+Validation loss decreased (25.691083 --> 24.331222).  Saving model ...
+
+Epoch: 8 	Training Loss: 22.215979 	Validation Loss: 23.632853
+Validation loss decreased (24.331222 --> 23.632853).  Saving model ...
+
+Epoch: 9 	Training Loss: 21.408623 	Validation Loss: 23.475442
+Validation loss decreased (23.632853 --> 23.475442).  Saving model ...
+
+Epoch: 10 	Training Loss: 20.637072 	Validation Loss: 23.639358
+
+Epoch: 11 	Training Loss: 19.877338 	Validation Loss: 22.408472
+Validation loss decreased (23.475442 --> 22.408472).  Saving model ...
+
+Epoch: 12 	Training Loss: 19.188079 	Validation Loss: 23.296445
+
+Epoch: 13 	Training Loss: 18.647543 	Validation Loss: 22.897815
+
+Epoch: 14 	Training Loss: 17.989626 	Validation Loss: 22.755968
+
+the performance is as follow: ![alt text](perf_model_1.png)
+
+and the final accuries were:
+![Accuracies](final_accuracy_first_model.png)
+
+
+SECOND MODEL:
+
+for the second model, the validation loss goes lower, thougth in addition to the architectural changes, there are also just more weigth and it is longer to train.
+
+we archieve a valisation loss of 16, and the model is still improving after a larger number of epoch ( 20 vs 10)
+
+
+here are the final accuracies:
+Test Loss: 16.123924
+
+Test Accuracy of airplane: 81% (810/1000)
+Test Accuracy of automobile: 85% (855/1000)
+Test Accuracy of  bird: 63% (633/1000)
+Test Accuracy of   cat: 52% (525/1000)
+Test Accuracy of  deer: 69% (695/1000)
+Test Accuracy of   dog: 71% (717/1000)
+Test Accuracy of  frog: 77% (772/1000)
+Test Accuracy of horse: 77% (772/1000)
+Test Accuracy of  ship: 84% (843/1000)
+Test Accuracy of truck: 76% (765/1000)
+
+Test Accuracy (Overall): 73% (7387/10000)
+
 %% 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")
 ```

 %% 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")
 ```

 %% 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:code id:1458a562 tags:
+
+``` python
+model = Net_3Conv_3lin_Quant()
+print(model)
+
+
+#prepare for fusion
+model.eval()
+#model.qconfig = torch.ao.quantization.get_default_qat_config('86')
+#model_fused = torch.ao.quantization.fuse_modules(model,[['conv', 'relu',]])
+
+model.train()
+model_prepared = torch.ao.quantization.prepare_qat(model)
+training_loop(model_prepared)
+model_prepared.eval()
+model_quantized = torch.ao.quantization.convert(model_prepared)
+
+evaluate(model_quantized)
+print_size_of_model(model_quantized)
+
+```
+
 %% 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: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.

--- a/final_accuracy_first_model.png
+++ b/final_accuracy_first_model.png
--- a/perf_model_1.png
+++ b/perf_model_1.png
No results found