... requires_grad=False) # Create random Tensors for weights, and wrap them in ... w2 respectively. loss.backward() # Update weights using gradient descent; ...
確定! 回上一頁