Morten Hjorth-Jensen
11 lines
251 B
Python
11 lines
251 B
Python
import torch
|
|
w = torch.tensor(1.0, requires_grad = True)
|
|
b = torch.tensor(0.5, requires_grad = True)
|
|
x = torch.tensor([1.4])
|
|
y = torch.tensor([2.1])
|
|
z = torch.add(torch.mul(w,x),b)
|
|
loss = (y-z).pow(2).sum()
|
|
loss.backward()
|
|
print(w.grad)
|
|
print(b.grad)
|