The meaning of reduce=False


loss = nn.MSELoss(reduce=False)
loss = loss.mean()


loss = nn.MSELoss(reduce=True)

The same?
If yes, is it the same for all loss functions with the reduce option? If no, what do I need to do to get the same result if I choose to do reduce=False?
At line 1553 you can see the code for a pointwise loss. Then yes, it is the same if size_average is true.