oneflow.nn.utils.rnn.PackedSequence¶

class oneflow.nn.utils.rnn.PackedSequence(data: oneflow.Tensor, batch_sizes: Optional[oneflow.Tensor] = None, sorted_indices: Optional[oneflow.Tensor] = None, unsorted_indices: Optional[oneflow.Tensor] = None)¶

The interface is consistent with PyTorch. The documentation is referenced from: https://pytorch.org/docs/1.10/generated/torch.nn.utils.rnn.PackedSequence.html.

Holds the data and list of batch_sizes of a packed sequence.

All RNN modules accept packed sequences as inputs.

Note

Instances of this class should never be created manually. They are meant to be instantiated by functions like pack_padded_sequence().

Batch sizes represent the number elements at each sequence step in the batch, not the varying sequence lengths passed to pack_padded_sequence(). For instance, given data abc and x the PackedSequence would contain data axbc with batch_sizes=[2,1,1].

data¶

Tensor containing packed sequence

Type: Tensor

batch_sizes¶

Tensor of integers holding information about the batch size at each sequence step

Type: Tensor

sorted_indices¶

Tensor of integers holding how this PackedSequence is constructed from sequences.

Type: Tensor, optional

unsorted_indices¶

Tensor of integers holding how this to recover the original sequences with correct order.

Type: Tensor, optional

Note

data can be on arbitrary device and of arbitrary dtype. sorted_indices and unsorted_indices must be oneflow.int64 tensors on the same device as data.

However, batch_sizes should always be a CPU oneflow.int64 tensor.

This invariant is maintained throughout PackedSequence class, and all functions that construct a :class:PackedSequence in PyTorch (i.e., they only pass in tensors conforming to this constraint).