ナビゲーションバー・ドロップダウンメニューのHTML+CSSサンプルコード

    optimizer.zero_grad() to clear the past gradients.
    loss.backward() back-propagates, computing the current gradient.
    optimizer.step() updates the network parameters based on the gradient

or this case
    for i in range(num):
        loss+=Loss(input,target)
    optimizer.zero_grad() clears the past gradients.
    loss.backward() back-propagate and compute the current gradient.
    optimizer.step() update the network parameters according to the gradient

.backward()

    optimizer.zero_grad() clears the previous gradient.
    loss1.backward(retain_graph=True) backpropagate and compute the current gradient.
    loss2.backward(retain_graph=True) backward propagation, calculate current gradient.
    optimizer.step() updates the network parameters according to the gradient