Fix Gradients accumulation example

[dobosevych]

Link: https://pytorch.org/ignite/faq.html
It should be

accumulation_steps = 4

def update_fn(engine, batch):
    model.train()

    x, y = prepare_batch(batch, device=device, non_blocking=non_blocking)
    y_pred = model(x)
    loss = criterion(y_pred, y) / accumulation_steps
    loss.backward()

    if engine.state.iteration % accumulation_steps == 0:
        optimizer.step()
        optimizer.zero_grad()

    return loss.item()

trainer = Engine(update_fn)

instead of

accumulation_steps = 4

def update_fn(engine, batch):
    model.train()

    if engine.state.iteration % accumulation_steps == 0:
        optimizer.zero_grad()

    x, y = prepare_batch(batch, device=device, non_blocking=non_blocking)
    y_pred = model(x)
    loss = criterion(y_pred, y) / accumulation_steps
    loss.backward()

    if engine.state.iteration % accumulation_steps == 0:
        optimizer.step()

    return loss.item()

trainer = Engine(update_fn)

[vfdev-5]

@dobosevych thanks! yes, there is a bug in FAQ with optimizer.step(), it should be :

    if (engine.state.iteration + 1) % accumulation_steps == 0:
        optimizer.step()

Your code is shorter and more neat, I'll change the FAQ to this.