torch_brain.data#

torch_brain.data.collate#

`chain`	Wrap an object to specify that it (or any of its members) should be stacked along the first dimension when batching.
`collate`	Extension of PyTorch's `default_collate` function to enable more advanced collation of samples of variable lengths.
`pad`	Wrap an object to specify that it (or any of its members) should be padded to the maximum length in the batch.
`pad8`	Wrap an object to specify that it (or any of its members) should be padded to the maximum length in the batch.
`pad2d`
`pad2d8`
`track_batch`	Wrap an array or tensor to track the batch_index.
`track_mask`	Wrap an array or tensor to specify that its padding mask should be tracked.
`track_mask8`	Wrap an array or tensor to specify that its padding mask should be tracked.
`track_mask2d`	Wrap an array or tensor to specify that its padding mask should be tracked.
`track_mask2d8`	Wrap an array or tensor to specify that its padding mask should be tracked.

Deprecated.