# ============================================================
# MNIST Digit Classification using a Simple Neural Network (PyTorch)
# ============================================================

import torch
import torch.nn as nn
import torch.optim as optim
from torch.utils.data import DataLoader
import torchvision
import torchvision.transforms as transforms
import matplotlib.pyplot as plt
import numpy as np

# ------------------------------------------------------------
# 1. Device Configuration (GPU if available)
# ------------------------------------------------------------
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
print("Using device:", device)

# ------------------------------------------------------------
# 2. Load MNIST Dataset
# ------------------------------------------------------------
transform = transforms.Compose([
    transforms.ToTensor(),                # convert to tensor
    transforms.Normalize((0.5,), (0.5,))  # normalize
])

train_dataset = torchvision.datasets.MNIST(
    root='./data', train=True, transform=transform, download=True)

test_dataset = torchvision.datasets.MNIST(
    root='./data', train=False, transform=transform, download=True)

train_loader = DataLoader(train_dataset, batch_size=64, shuffle=True)
test_loader = DataLoader(test_dataset, batch_size=64, shuffle=False)

# ------------------------------------------------------------
# 3. Define Simple Neural Network
# ------------------------------------------------------------
class SimpleNN(nn.Module):
    def __init__(self):
        super(SimpleNN, self).__init__()
        self.fc1 = nn.Linear(28*28, 128)
        self.fc2 = nn.Linear(128, 64)
        self.fc3 = nn.Linear(64, 10)
        self.relu = nn.ReLU()

    def forward(self, x):
        x = x.view(-1, 784)    # flatten image
        x = self.relu(self.fc1(x))
        x = self.relu(self.fc2(x))
        x = self.fc3(x)        # no softmax (CrossEntropyLoss handles it)
        return x

model = SimpleNN().to(device)

# ------------------------------------------------------------
# 4. Loss and Optimizer
# ------------------------------------------------------------
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(model.parameters(), lr=0.001)

# ------------------------------------------------------------
# 5. Training Loop
# ------------------------------------------------------------
num_epochs = 5
train_loss_list = []
train_acc_list = []

for epoch in range(num_epochs):
    total_loss = 0
    correct = 0
    total = 0

    for images, labels in train_loader:
        images, labels = images.to(device), labels.to(device)

        # forward pass
        outputs = model(images)
        loss = criterion(outputs, labels)

        # backward pass
        optimizer.zero_grad()
        loss.backward()
        optimizer.step()

        total_loss += loss.item()

        # accuracy
        _, predicted = outputs.max(1)
        total += labels.size(0)
        correct += (predicted == labels).sum().item()

    avg_loss = total_loss / len(train_loader)
    accuracy = 100 * correct / total

    train_loss_list.append(avg_loss)
    train_acc_list.append(accuracy)

    print(f"Epoch [{epoch+1}/{num_epochs}]  Loss: {avg_loss:.4f}  Accuracy: {accuracy:.2f}%")

# ------------------------------------------------------------
# 6. Testing / Evaluation
# ------------------------------------------------------------
model.eval()
correct = 0
total = 0

with torch.no_grad():
    for images, labels in test_loader:
        images, labels = images.to(device), labels.to(device)
        outputs = model(images)

        _, predicted = outputs.max(1)
        total += labels.size(0)
        correct += (predicted == labels).sum().item()

test_accuracy = 100 * correct / total
print("\nTest Accuracy:", test_accuracy)

# ------------------------------------------------------------
# 7. Plot Accuracy and Loss Graphs
# ------------------------------------------------------------
plt.figure(figsize=(8, 4))
plt.plot(train_acc_list, label="Training Accuracy")
plt.title("Training Accuracy")
plt.xlabel("Epoch")
plt.ylabel("Accuracy (%)")
plt.grid()
plt.legend()
plt.show()

plt.figure(figsize=(8, 4))
plt.plot(train_loss_list, label="Training Loss")
plt.title("Training Loss")
plt.xlabel("Epoch")
plt.ylabel("Loss")
plt.grid()
plt.legend()
plt.show()

# ------------------------------------------------------------
# 8. Show Predictions for First 5 Test Images
# ------------------------------------------------------------
data_iter = iter(test_loader)
images, labels = next(data_iter)
images, labels = images.to(device), labels.to(device)

outputs = model(images[:5])
_, preds = outputs.max(1)

print("\nPredicted:", preds.cpu().numpy())
print("Actual:   ", labels[:5].cpu().numpy())

# Plot the first 5 images
plt.figure(figsize=(10, 2))
for i in range(5):
    plt.subplot(1, 5, i + 1)
    plt.imshow(images[i].cpu().squeeze(), cmap="gray")
    plt.title(f"P:{preds[i].item()} / T:{labels[i].item()}")
    plt.axis("off")

plt.show()

-------------------------------------------------------------------------------------------------
//Image dataset

import torch
import torch.nn as nn
import torch.optim as optim
from torch.utils.data import DataLoader, random_split
from torchvision import datasets, transforms
import matplotlib.pyplot as plt
import numpy as np  # Added for image visualization
import os

# -----------------------------------------------------------
# 1. Device Configuration
# -----------------------------------------------------------
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
print("Using device:", device)

# -----------------------------------------------------------
# 2. Dataset Path
# -----------------------------------------------------------
# UPDATE THIS PATH to point to your actual dataset folder
dataset_dir = r"C:\Users\LOMESH\Downloads\archive\natural_images"

# Check if path exists to avoid confusing errors later
if not os.path.exists(dataset_dir):
    print(f"ERROR: The path '{dataset_dir}' does not exist.")
    print("Please update the 'dataset_dir' variable.")
else:
    # -----------------------------------------------------------
    # 3. Transformations (preprocessing + augmentation)
    # -----------------------------------------------------------
    transform = transforms.Compose([
        transforms.Resize((128, 128)),
        transforms.RandomHorizontalFlip(),
        transforms.ToTensor(),
        transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
    ])

    # Load the full dataset
    full_dataset = datasets.ImageFolder(root=dataset_dir, transform=transform)

    # Split dataset into training and validation (80%-20%)
    train_size = int(0.8 * len(full_dataset))
    val_size = len(full_dataset) - train_size
    train_dataset, val_dataset = random_split(full_dataset, [train_size, val_size])

    train_loader = DataLoader(train_dataset, batch_size=32, shuffle=True)
    val_loader = DataLoader(val_dataset, batch_size=32, shuffle=False)

    num_classes = len(full_dataset.classes)
    print("Classes:", full_dataset.classes)

    # -----------------------------------------------------------
    # 4. Define CNN Model
    # -----------------------------------------------------------
    class SimpleCNN(nn.Module):
        def __init__(self, num_classes):
            super(SimpleCNN, self).__init__()
            self.conv_layers = nn.Sequential(
                nn.Conv2d(3, 32, kernel_size=3, padding=1),
                nn.ReLU(),
                nn.MaxPool2d(2, 2),
                
                nn.Conv2d(32, 64, kernel_size=3, padding=1),
                nn.ReLU(),
                nn.MaxPool2d(2, 2),
                
                nn.Conv2d(64, 128, kernel_size=3, padding=1),
                nn.ReLU(),
                nn.MaxPool2d(2, 2)
            )
            
            # Calculate flattened size: 128 -> 64 -> 32 -> 16 (after 3 pools)
            # 128 channels * 16 * 16 spatial size
            self.fc_layers = nn.Sequential(
                nn.Flatten(),
                nn.Linear(128 * 16 * 16, 256),
                nn.ReLU(),
                nn.Linear(256, num_classes)
            )

        def forward(self, x):
            x = self.conv_layers(x)
            x = self.fc_layers(x)
            return x

    model = SimpleCNN(num_classes=num_classes).to(device)

    # -----------------------------------------------------------
    # 5. Loss and Optimizer
    # -----------------------------------------------------------
    criterion = nn.CrossEntropyLoss()
    optimizer = optim.Adam(model.parameters(), lr=0.001)

    # -----------------------------------------------------------
    # 6. Training Loop
    # -----------------------------------------------------------
    num_epochs = 10
    train_loss_list, train_acc_list = [], []

    print("Starting Training...")
    for epoch in range(num_epochs):
        model.train()
        total_loss, correct, total = 0, 0, 0
        
        for images, labels in train_loader:
            images, labels = images.to(device), labels.to(device)
            
            # Forward pass
            outputs = model(images)
            loss = criterion(outputs, labels)
            
            # Backward pass
            optimizer.zero_grad()
            loss.backward()
            optimizer.step()
            
            total_loss += loss.item()
            _, predicted = outputs.max(1)
            total += labels.size(0)
            correct += (predicted == labels).sum().item()
        
        avg_loss = total_loss / len(train_loader)
        accuracy = 100 * correct / total
        train_loss_list.append(avg_loss)
        train_acc_list.append(accuracy)
        
        print(f"Epoch [{epoch+1}/{num_epochs}]  Loss: {avg_loss:.4f} Accuracy: {accuracy:.2f}%")

    # -----------------------------------------------------------
    # 7. Validation / Evaluation
    # -----------------------------------------------------------
    model.eval()
    correct, total = 0, 0
    with torch.no_grad():
        for images, labels in val_loader:
            images, labels = images.to(device), labels.to(device)
            outputs = model(images)
            _, predicted = outputs.max(1)
            total += labels.size(0)
            correct += (predicted == labels).sum().item()

    val_accuracy = 100 * correct / total
    print(f"\nValidation Accuracy: {val_accuracy:.2f}%")

    # -----------------------------------------------------------
    # 8. Plot Training Accuracy and Loss
    # -----------------------------------------------------------
    plt.figure(figsize=(10, 4))
    
    # Plot Accuracy
    plt.subplot(1, 2, 1)
    plt.plot(train_acc_list, label="Training Accuracy", color='blue')
    plt.title("Training Accuracy")
    plt.xlabel("Epoch")
    plt.ylabel("Accuracy (%)")
    plt.grid()
    plt.legend()

    # Plot Loss
    plt.subplot(1, 2, 2)
    plt.plot(train_loss_list, label="Training Loss", color='red')
    plt.title("Training Loss")
    plt.xlabel("Epoch")
    plt.ylabel("Loss")
    plt.grid()
    plt.legend()

    plt.tight_layout()
    plt.show()

    # -----------------------------------------------------------
    # 9. Show Predictions for First 5 Validation Images
    # -----------------------------------------------------------
    data_iter = iter(val_loader)
    images, labels = next(data_iter)
    images, labels = images.to(device), labels.to(device)

    # Get predictions
    outputs = model(images[:5])
    _, preds = outputs.max(1)

    print("\nPredicted:", [full_dataset.classes[i] for i in preds.cpu().numpy()])
    print("Actual:   ", [full_dataset.classes[i] for i in labels[:5].cpu().numpy()])

    plt.figure(figsize=(12, 3))
    for i in range(5):
        plt.subplot(1, 5, i+1)
        
        # Un-normalize image for display
        img = images[i].cpu().permute(1, 2, 0).numpy()
        mean = np.array([0.485, 0.456, 0.406])
        std = np.array([0.229, 0.224, 0.225])
        img = std * img + mean
        img = np.clip(img, 0, 1) # Clip values to be between 0 and 1
        
        plt.imshow(img)
        plt.title(f"P:{full_dataset.classes[preds[i]]}\nT:{full_dataset.classes[labels[i]]}", fontsize=10)
        plt.axis("off")
        
    plt.show()
-----------------------------------------------------------------------------
//Iris dataset

import torch
import torch.nn as nn
import torch.optim as optim
from torch.utils.data import DataLoader, TensorDataset, random_split
from sklearn.datasets import load_iris
import matplotlib.pyplot as plt

# ---------------------------------------------------------
# NOTE: If this crashes with "ImportError: numpy.core...", 
# run this in a separate cell first: !pip install "numpy<2"
# ---------------------------------------------------------

device = 'cuda' if torch.cuda.is_available() else 'cpu'
print(f"Using device: {device}")

# ----- DATA -----
# Load data
X, y = load_iris(return_X_y=True)

# Convert to PyTorch tensors
X = torch.tensor(X, dtype=torch.float32)
y = torch.tensor(y, dtype=torch.long)

# Normalize (Standardization)
# Note: In real projects, calculate mean/std on train_ds only to avoid leakage
X = (X - X.mean(0)) / X.std(0)

# Split 120 train / 30 test
train_ds, test_ds = random_split(TensorDataset(X, y), [120, 30])

train_loader = DataLoader(train_ds, batch_size=16, shuffle=True)
test_loader  = DataLoader(test_ds, batch_size=16)

# ----- MODEL -----
model = nn.Sequential(
    nn.Linear(4, 32),
    nn.ReLU(),
    nn.Linear(32, 3)
).to(device)

loss_fn = nn.CrossEntropyLoss()
opt = optim.Adam(model.parameters(), lr=0.01)

# ----- TRAIN -----
loss_hist, acc_hist = [], []
print("Starting training...")

for epoch in range(50):
    correct = 0
    total = 0
    running_loss = 0.0
    
    for x, y_batch in train_loader:
        x, y_batch = x.to(device), y_batch.to(device)
        
        # Forward
        out = model(x)
        loss = loss_fn(out, y_batch)
        
        # Backward
        opt.zero_grad()
        loss.backward()
        opt.step()
        
        # Metrics
        running_loss += loss.item()
        correct += (out.argmax(1) == y_batch).sum().item()
        total += y_batch.size(0)

    # Store epoch metrics
    epoch_loss = running_loss / len(train_loader)
    epoch_acc = 100 * correct / total
    loss_hist.append(epoch_loss)
    acc_hist.append(epoch_acc)

print(f"Final Train Accuracy: {acc_hist[-1]:.2f}%")

# ----- TEST -----
model.eval() # Good practice to set eval mode
correct = 0
total = 0

with torch.no_grad():
    for x, y_batch in test_loader:
        x, y_batch = x.to(device), y_batch.to(device)
        correct += (model(x).argmax(1) == y_batch).sum().item()
        total += y_batch.size(0)

print(f"Test Accuracy: {100 * correct / total:.2f}%")

# ----- PLOT -----
plt.figure(figsize=(10, 4))
plt.subplot(1, 2, 1)
plt.plot(loss_hist)
plt.title("Loss")
plt.xlabel("Epoch")

plt.subplot(1, 2, 2)
plt.plot(acc_hist)
plt.title("Accuracy")
plt.xlabel("Epoch")

plt.show()