cnn.py

#%% packages
import numpy as np
from sklearn.metrics import accuracy_score
import torch

import torchvision

import torchvision.transforms as transforms
import torch.nn as nn
# %%
transform_super = transforms.Compose(
    [transforms.Resize(32),
    transforms.Grayscale(num_output_channels=1),
    transforms.ToTensor(),
    transforms.Normalize((0.5, ), (0.5, ))])

batch_size = 4
trainset = torchvision.datasets.ImageFolder(root='data/train', transform=transform_super)
trainloader = torch.utils.data.DataLoader(trainset, batch_size=batch_size, shuffle=True)
testset = torchvision.datasets.ImageFolder(root='data/test', transform=transform_super)
testloader = torch.utils.data.DataLoader(testset, batch_size=batch_size, shuffle=True)
classes = ['healthy', 'sick']

# %% Neural Network setup
class ImageClassificationNet(nn.Module):
    def __init__(self) -> None:
        super().__init__()
        self.conv1 = nn.Conv2d(1, 6, 3)
        self.pool = nn.MaxPool2d(2, 2)
        self.conv2 = nn.Conv2d(6, 16, 3)
        self.fc1 = nn.Linear(16 * 6 * 6, 128)
        self.fc2 = nn.Linear(128, 64)
        self.fc3 = nn.Linear(64, 1)
        self.relu = nn.ReLU()
        self.sigmoid = nn.Sigmoid()
    
    def forward(self, x):
        x = self.conv1(x)
        x = self.relu(x)
        x = self.pool(x)
        x = self.conv2(x)
        x = self.relu(x)
        x = self.pool(x)
        x = torch.flatten(x, 1)
        x = self.fc1(x)
        x = self.relu(x)
        x = self.fc2(x)
        x = self.relu(x)
        x = self.fc3(x)
        x = self.sigmoid(x)
        return x

#%% init model
model = ImageClassificationNet()      
loss_fn = nn.BCELoss()
optimizer = torch.optim.SGD(model.parameters(), lr=0.001, momentum=0.8)
# %% training
NUM_EPOCHS = 10
losses = []
for epoch in range(NUM_EPOCHS):
    epoch_loss = 0
    for i, data in enumerate(trainloader, 0):
        print(f'Epoch {epoch}/{NUM_EPOCHS}, Step {i+1}/{len(trainloader)}')
        inputs, labels = data
        optimizer.zero_grad()
        outputs = model(inputs)
        
        loss = loss_fn(outputs, labels.reshape(-1, 1).float())
        loss.backward()
        optimizer.step()
        # add loss to epoch loss
        epoch_loss += loss.item()
    losses.append(epoch_loss)
            
#%% losses
import seaborn as sns
sns.lineplot(x=range(NUM_EPOCHS), y=losses)

# %% test
y_test = []
y_test_hat = []
for i, data in enumerate(testloader, 0):
    inputs, y_test_temp = data
    with torch.no_grad():
        y_test_hat_temp = model(inputs).round()
    
    y_test.extend(y_test_temp.numpy())
    y_test_hat.extend(y_test_hat_temp.numpy())

# %%
acc = accuracy_score(y_test, y_test_hat)
print(f'Accuracy: {acc*100:.2f} %')
# %%
# We know that data is balanced, so baseline classifier has accuracy of 50 %.