BatchIterator

class BatchIterator(batch_size: int = 32, shuffle: bool = True)

Bases: bluebird.data.DataIterator

Default iterator, returns batches of input, target pairs.

Batches are randomized by default.

Example:

iterator = BatchIterator()
for batch in iterator(Data):
    # do something

__call__(inputs: Tensor, targets: Tensor) -> Iterator[Batch]:

Returns batches of data.

Args:

inputs (Tensor): input to the network targets (Tensor): expected value

Returns:

Iterator[Batch]: Batches of data

__init__(batch_size: int = 32, shuffle: bool = True) → None

Initalizes the object.

Parameters

• batch_size (int) – length of every batch size, defaults to 32

• shuffle (bool) – shuffles data if true, defaults to True

Methods

__init__([batch_size, shuffle])

Initalizes the object.