zero.stream.Stream.epochs¶

Stream.epochs(n_epochs, epoch_size=None, progress_bar=True)[source]¶

Iterate over data epochs.

A shortcut for what is probably the most popular form of a training loop in Deep Learning (plus a progress bar):

for epoch in stream.epochs(n_epochs, epoch_size):
    for x in epoch:
        ...

# is equivalent to:

while stream.epoch < n_epochs:
    stream.increment_epoch()
    for x in stream.data(epoch_size):
        ...

Parameters

n_epochs (Union[int, float]) – the number of epochs. If float, must be math.inf.
epoch_size (Optional[Union[int, float]]) – the number of data items in one epoch (is forwarded to Stream.data)
progress_bar (bool) – show the progress bar for iterations. The initial value is set to Stream.iteration. See also the note below.

Returns

Iterator over iterators over data from Stream.loader.

Raises

AssertionError – if n_epochs if float, but not math.inf.

Return type

Iterator[Iterator[Any]]

Note

If progress_bar is True, the progress bar is updated on yielding every item which means that the progress bar should be interpreted as “what iteration is in progress” instead of “how many iterations are done”. The percentage will be displayed only if the total number of planned iterations can be inferred from the arguments and/or from Stream.loader.

Examples

stream = Stream(range(3))
for epoch in stream.epochs(2):
    for x in epoch:
        print(x)
    print('-')

stream = Stream(range(3))
for epoch in stream.epochs(3, 2):
    for x in epoch:
        print(x)
    print('-')