Cant find solution to error - Model 2: Training Our First CNN and Evaluating Its Results

[tecdungeon]

Error is:

Traceback (most recent call last): File "D:\TWSAPI\source\pythonclient\tws_env\Code\PyTorch_Vision.py", line 308, in train_step(model=model_2, File "D:\TWSAPI\source\pythonclient\tws_env\Code\PyTorch_Vision.py", line 158, in train_step y_pred = model(X) ^^^^^^^^ File "D:\TWSAPI\source\pythonclient.venv\Lib\site-packages\torch\nn\modules\module.py", line 1511, in _wrapped_call_impl return self._call_impl(*args, *kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "D:\TWSAPI\source\pythonclient.venv\Lib\site-packages\torch\nn\modules\module.py", line 1520, in _call_impl return forward_call(args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "D:\TWSAPI\source\pythonclient\tws_env\Code\PyTorch_Vision.py", line 115, in forward x - self.classifier(x) ^~~~~~ RuntimeError: The size of tensor a (7) must match the size of tensor b (10) at non-singleton dimension 3

I am assuming the main error is: RuntimeError: The size of tensor a (7) must match the size of tensor b (10) at non-singleton dimension 3 and its within the FashionMNISTModelV2 class (self.classifier)

Code:

import torch
from torch import nn
import torchvision
from torchvision import datasets
from torchvision import transforms
from torchvision.transforms import ToTensor
from torch.utils.data import DataLoader
from timeit import default_timer as timer
import matplotlib.pyplot as plt
import argparse
from tqdm.auto import tqdm
from pathlib import Path

if Path("helper_functions.py").is_file():
    print("helper_functions.py already exists")
else:
    print("downloading helper_functions")
    request = requests.get("https://raw.githubusercontent.com/mrdbourke/pytorch-deep-learning/main/helper_functions.py")
    with open("helper_functions.py", "wb") as f:
        f.write(request.content)

from helper_functions import plot_predictions, plot_decision_boundary,accuracy_fn

parser = argparse.ArgumentParser(description='CPU/GPU')
parser.add_argument('--disable-cuda', action='store_true', help='Disable CUDA')
args = parser.parse_args()
args.device = None

if not args.disable_cuda and torch.cuda.is_available():
    args.device = torch.device('cuda')
else:
    args.device = torch.device('cpu')

class FashionMNISTModelV0(nn.Module):
    def __init__(self,
                 input_shape: int,
                 hidden_units: int,
                 output_shape: int):
        super().__init__()
        self.layer_stack = nn.Sequential(
            nn.Flatten(),
            nn.Linear(in_features=input_shape,
                      out_features=hidden_units),
            nn.Linear(in_features=hidden_units,
                      out_features=output_shape),

        )
    def forward(self, x):
        return self.layer_stack(x)

class FashionMNISTModelV1(nn.Module):
    def __init__(self,
                 input_shape: int,
                 hidden_units: int,
                 output_shape: int):
        super().__init__()
        self.layer_stack = nn.Sequential(
            nn.Flatten(),
            nn.Linear(in_features=input_shape,
                      out_features=hidden_units),
            nn.ReLU(),
            nn.Linear(in_features=hidden_units,
                      out_features=output_shape),
            nn.ReLU(),
        )
    def forward(self, x):
        return self.layer_stack(x)

class FashionMNISTModelV2(nn.Module):
    def __init__(self, input_shape: int, hidden_units: int, output_shape: int):
        super().__init__()
        self.conv_block_1 = nn.Sequential(
            nn.Conv2d(in_channels=input_shape,
                      out_channels=hidden_units,
                      kernel_size=3,
                      stride=1,
                      padding=1),
            nn.ReLU(),
            nn.Conv2d(in_channels=hidden_units,
                      out_channels=hidden_units,
                      kernel_size=3,
                      stride=1,
                      padding=1),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2,
                         stride=2)
        )
        self.conv_block_2 = nn.Sequential(
            nn.Conv2d(in_channels=hidden_units,
                      out_channels=hidden_units,
                      kernel_size=3,
                      stride=1,
                      padding=1),
            nn.ReLU(),
            nn.Conv2d(in_channels=hidden_units,
                      out_channels=hidden_units,
                      kernel_size=3,
                      stride=1,
                      padding=1),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2)
        )
        self.classifier = nn.Sequential(
            nn.Flatten(),
            nn.Linear(in_features=10*7*7,
                      out_features=output_shape)
        )

    def forward(self, x):
        print(f"pre conv 1 {x.shape}")
        x = self.conv_block_1(x)
        print(f"conv 1 {x.shape}")
        x = self.conv_block_2(x)
        print(f"conv 2 {x.shape}")
        x - self.classifier(x)
        print(x.shape)
        return x

def print_train_time(start: float,
                     end: float,
                     device: torch.device = None):
    total_time = end - start
    print(f'Training time on {device}: {total_time:.3f} seconds')
    return total_time

def eval_model(model: torch.nn.Module,
               data_loader: torch.utils.data.DataLoader,
               loss_fn: torch.nn.Module,
               accuracy_fn,
               device=args.device):
    loss, acc = 0,0
    model.eval()
    with torch.inference_mode():
        for X, y in tqdm(data_loader):
            X,y = X.to(device), y.to(device)
            y_pred = model(X)
            loss += loss_fn(y_pred, y)
            acc += accuracy_fn(y_true=y,
                               y_pred=y_pred.argmax(dim=1))
        loss = loss / len(data_loader)
        acc = acc / len(data_loader)
    return {"model_name": model.__class__.__name__,
            "model_loss": loss.item(),
            "model_acc": acc}

def train_step(model: torch.nn.Module,
               data_loader: torch.utils.data.DataLoader,
               loss_fn: torch.nn.Module,
               optimizer: torch.optim.Optimizer,
               accuracy_fn,
               device: torch.device = args.device):
    train_loss, train_acc = 0, 0
    for batch, (X, y) in enumerate(data_loader):
        # Send data to GPU
        X, y = X.to(device), y.to(device)

        # 1. Forward pass
        y_pred = model(X)

        # 2. Calculate loss
        loss = loss_fn(y_pred, y)
        train_loss += loss
        train_acc += accuracy_fn(y_true=y,
                                 y_pred=y_pred.argmax(dim=1)) # Go from logits -> pred labels

        # 3. Optimizer zero grad
        optimizer.zero_grad()

        # 4. Loss backward
        loss.backward()

        # 5. Optimizer step
        optimizer.step()

    # Calculate loss and accuracy per epoch and print out what's happening
    train_loss /= len(data_loader)
    train_acc /= len(data_loader)
    print(f"Train loss: {train_loss:.5f} | Train accuracy: {train_acc:.2f}%")

def test_step(model: torch.nn.Module,
              data_loader: torch.utils.data.DataLoader,
              loss_fn: torch.nn.Module,
              accuracy_fn,
              device: torch.device = args.device):
    test_loss,test_acc = 0, 0
    model.eval()
    with torch.inference_mode():
        for X, y in test_dataloader:
            X, y = X.to(device), y.to(device)
            test_pred = model(X)
            test_loss += loss_fn(test_pred, y)
            test_acc += accuracy_fn(y_true=y, y_pred=test_pred.argmax(dim=1))

        test_loss /= len(data_loader)
        test_acc /= len(data_loader)
        print(f"Test Loss: {test_loss:.5f} | Test Accuracy: {test_acc:.5f}%\n")

torch.manual_seed(42)
train_data = datasets.FashionMNIST('../data',
                            train=True,
                            download=True,
                            transform=torchvision.transforms.ToTensor(),
                            target_transform=None)

test_data = datasets.FashionMNIST('../data',
                            train=False,
                            download=True,
                            transform=torchvision.transforms.ToTensor(),
                            target_transform=None)

BATCH_SIZE = 32

train_dataloader = DataLoader(dataset=train_data,
                              batch_size=BATCH_SIZE,
                              shuffle=True)

test_dataloader = DataLoader(dataset=test_data,
                             batch_size=BATCH_SIZE,
                             shuffle=False)

image, label = train_data[0]
class_to_idx = train_data.class_to_idx
class_names = train_data.classes
# print(f"Image shape: {image.shape}"
#       f"{class_to_idx}")
#
# fig=plt.figure(figsize=(9,9))
# rows, cols = 4, 4
# for i in range(1,rows*cols+1):
#     random_idx = torch.randint(low=0, high=len(train_data), size=[1]).item()
#     img, label = train_data[random_idx]
#     fig.add_subplot(rows, cols,i)
#     print(f"random_idx: {random_idx} label: {label} img shape: {img.shape}")
#
#     plt.imshow(img.squeeze(), cmap="gray")
#     plt.title(class_names[label])
#     plt.axis(False)
#     #print(random_idx)
# # plt.imshow(image.squeeze(),cmap="gray")
# # plt.title(class_names[label])
# # plt.axis(False)
# plt.show()

train_features_batch, train_labels_batch = next(iter(train_dataloader))
flatten_model = nn.Flatten()

x = train_features_batch[0]
output = flatten_model(x)

train_time_start = timer()
model_0 = FashionMNISTModelV0(
    input_shape=28*28,
    hidden_units=10,
    output_shape=len(class_names)
).to("cpu")

train_time_start = timer()
model_1 = FashionMNISTModelV1(
    input_shape=28*28,
    hidden_units=10,
    output_shape=len(class_names)
).to(args.device)

model_2 = FashionMNISTModelV2(
    input_shape=1,
    hidden_units=10,
    output_shape=len(class_names)
).to(args.device)

torch.manual_seed(42)

# Create sample batch of random numbers with same size as image batch
images = torch.randn(size=(32, 3, 64, 64)) # [batch_size, color_channels, height, width]
test_image = images[0] # get a single image for testing

torch.manual_seed(42)

# Create a convolutional layer with same dimensions as TinyVGG
# (try changing any of the parameters and see what happens)
conv_layer = nn.Conv2d(in_channels=3,
                       out_channels=10,
                       kernel_size=3,
                       stride=1,
                       padding=0) # also try using "valid" or "same" here

# Pass the data through the convolutional layer
#print(conv_layer(test_image)) # Note: If running PyTorch <1.11.0, this will error because of shape issues (nn.Conv.2d() expects a 4d tensor as input)

conv_layer_2 = nn.Conv2d(in_channels=3, # same number of color channels as our input image
                         out_channels=10,
                         kernel_size=(5, 5), # kernel is usually a square so a tuple also works
                         stride=2,
                         padding=0)

loss_fn = nn.CrossEntropyLoss()
optimizer = torch.optim.SGD(params=model_2.parameters(),
            lr=0.1)

#print(str(next(model_1.parameters()).device))
epochs = 3

#train_time_start_on_cpu = timer()
train_time_start_on_gpu = timer()

for epoch in tqdm(range(epochs)):
    print(f"Epoch: {epoch}\n-------")
    train_step(model=model_2,
               data_loader=train_dataloader,
               loss_fn=loss_fn,
               optimizer=optimizer,
               accuracy_fn=accuracy_fn,
               device=args.device)

    test_step(model=model_2,
              data_loader=test_dataloader,
              loss_fn=loss_fn,
              accuracy_fn=accuracy_fn,
              device=args.device)

#train_time_end_on_cpu = timer()
train_time_end_on_gpu = timer()
total_train_time_model_1 = print_train_time(start=train_time_start_on_gpu,
                                            end=train_time_end_on_gpu,
                                            device=args.device)

model_1_results = eval_model(model=model_1,
                             data_loader=test_dataloader,
                             loss_fn=loss_fn,
                             accuracy_fn=accuracy_fn,
                             device=args.device)

print(model_1_results)

[tecdungeon]

Error was:

def forward(self, x):

Should have been:

def forward(self, x: torch.Tensor):