Backpropagation using Pytorch

backpropagation.py

import torch
'''
Backpropgation

eg. x = [1.,1.,1]  -> "This is the fixed feature value:"
    y = x + 2  -> a(x) -> "This is the respective y value:"
    z = y*y*3  -> b(y) -> "This is the example for loss function"
Backpropgation    x -> a(x) -> b(y) -> z

dz/dx = dz/dy * dy/dx (Chain Rule)

'''
# Find the gradient provided the respective x, y, w
x = torch.tensor(1.0)
y = torch.tensor(2.0)
w = torch.tensor(-2.0, requires_grad=True)

# forward path and compute loss
y_hat =w * x 
loss = (y_hat - y)**2
print(loss)

# backward pass

loss.backward()

print(w.grad)

## TODO: update weight
## TODO: next next forward and backward