Basic Tensor Operations

import torch
import numpy as np
import seaborn as sns

# Create a tensor x with a scalar value 4.5
x = torch.tensor(4.5)

# Define y as the result of adding 10 to x
y = x + 10

# Print the value of y
print(y)

# Check if x requires gradient tracking
print(x.requires_grad)

# Create a tensor z with a scalar value 2.0 and enable gradient tracking
z = torch.tensor(2.0, requires_grad=True)

# Check if z requires gradient tracking
print(z.requires_grad)

# Define a function to compute a quadratic polynomial
def func(val):
    return (val - 3) * (val - 2) * (val - 4)

# Generate a range of x values and compute y values using the function
x_range = np.linspace(0, 10, 101)
y_range = [func(i) for i in x_range]

# Plot the function using seaborn
sns.lineplot(x=x_range, y=y_range)

# Redefine z with a tensor of value 5.0 and enable gradient tracking
z = torch.tensor([5.0], requires_grad=True)

# Compute y as the result of a polynomial expression involving z
y = (z - 3) * (z - 2) * (z - 4)

# Perform backpropagation to compute gradients
y.backward()

# Print the gradient of z
print(z.grad)



# Second Example
# Define a more complex network with tensors and operations
x11 = torch.tensor(2.0, requires_grad=True)
x21 = torch.tensor(3.0, requires_grad=True)

# Define intermediate variables using operations on x11 and x21
x12 = 5 * x11 - 3 * x21
x22 = 2 * x11 ** 2 + 2 * x21

# Define y as a combination of x12 and x22
y = 4 * x12 + 3 * x22

# Perform backpropagation to compute gradients
y.backward()

# Print the gradients of x11 and x21
print(x11.grad)
print(x21.grad)