# PyTorch Tensor Fundamentals
import torch
import torch.nn as nn
import numpy as np

# Tensor Creation and Basic Operations
def tensor_basics():
    """Demonstrate PyTorch tensor creation and operations"""
    
    # Creating tensors
    x = torch.tensor([1, 2, 3, 4, 5], dtype=torch.float32)
    y = torch.zeros(3, 4)  # 3x4 tensor of zeros
    z = torch.randn(2, 3, 4)  # Random normal distribution
    
    # Device handling (CPU/GPU)
    device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
    x = x.to(device)
    
    print(f"Tensor x: {x}")
    print(f"Device: {x.device}")
    print(f"Shape: {x.shape}")
    print(f"Data type: {x.dtype}")
    
    # Mathematical operations
    a = torch.tensor([[1, 2], [3, 4]], dtype=torch.float32)
    b = torch.tensor([[5, 6], [7, 8]], dtype=torch.float32)
    
    # Element-wise operations
    add_result = a + b
    mul_result = a * b
    
    # Matrix operations
    matmul_result = torch.matmul(a, b)
    
    # Broadcasting
    scalar = 10
    broadcast_result = a + scalar
    
    return {
        'addition': add_result,
        'multiplication': mul_result,
        'matrix_multiplication': matmul_result,
        'broadcasting': broadcast_result
    }

# Advanced Tensor Operations
def advanced_tensor_operations():
    """Advanced tensor manipulation and operations"""
    
    # Reshaping and views
    x = torch.randn(4, 6)
    reshaped = x.view(2, 12)  # Reshape to 2x12
    flattened = x.flatten()   # Flatten to 1D
    
    # Indexing and slicing
    subset = x[0:2, 1:4]  # First 2 rows, columns 1-3
    
    # Concatenation and stacking
    a = torch.tensor([[1, 2], [3, 4]])
    b = torch.tensor([[5, 6], [7, 8]])
    
    concatenated = torch.cat([a, b], dim=0)  # Along rows
    stacked = torch.stack([a, b], dim=0)     # New dimension
    
    # Reduction operations
    sum_all = x.sum()
    sum_dim = x.sum(dim=1)  # Sum along columns
    mean_val = x.mean()
    max_val, max_idx = x.max(dim=0)
    
    # In-place operations (memory efficient)
    x.add_(1)  # Add 1 to all elements in-place
    
    return {
        'original_shape': x.shape,
        'reshaped_shape': reshaped.shape,
        'concatenated': concatenated,
        'sum': sum_all.item(),
        'mean': mean_val.item()
    }

# Example usage
if __name__ == "__main__":
    print("=== Tensor Basics ===")
    basics = tensor_basics()
    
    print("\n=== Advanced Operations ===")
    advanced = advanced_tensor_operations()