mindspore.ops.ReverseSequence

class mindspore.ops.ReverseSequence(*args, **kwargs)[source]

Reverses variable length slices.

Parameters
  • seq_dim (int) – The dimension where reversal is performed. Required.

  • batch_dim (int) – The input is sliced in this dimension. Default: 0.

Inputs:
  • x (Tensor) - The input to reverse, supporting all number types including bool.

  • seq_lengths (Tensor) - Must be a 1-D vector with int32 or int64 types.

Outputs:

Reversed tensor with the same shape and data type as input.

Raises

TypeError – If seq_dim or batch_dim is not an int.

Supported Platforms:

Ascend GPU

Examples

>>> x = Tensor(np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]]), mindspore.float32)
>>> seq_lengths = Tensor(np.array([1, 2, 3]))
>>> reverse_sequence = ops.ReverseSequence(seq_dim=1)
>>> output = reverse_sequence(x, seq_lengths)
>>> print(output)
[[1. 2. 3.]
 [5. 4. 6.]
 [9. 8. 7.]]
>>> x = Tensor(np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]]), mindspore.float32)
>>> seq_lengths = Tensor(np.array([1, 2, 3]))
>>> reverse_sequence = ops.ReverseSequence(seq_dim=0, batch_dim=1)
>>> output = reverse_sequence(x, seq_lengths)
>>> print(output)
[[1. 5. 9.]
 [4. 2. 6.]
 [7. 8. 3.]]
>>> x = Tensor(np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]]), mindspore.float32)
>>> seq_lengths = Tensor(np.array([2, 2, 3]))
>>> reverse_sequence = ops.ReverseSequence(seq_dim=1)
>>> output = reverse_sequence(x, seq_lengths)
>>> print(output)
[[2. 1. 3.]
 [5. 4. 6.]
 [9. 8. 7.]]
>>> x = Tensor(np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]]), mindspore.float32)
>>> seq_lengths = Tensor(np.array([3, 2, 3]))
>>> reverse_sequence = ops.ReverseSequence(seq_dim=1)
>>> output = reverse_sequence(x, seq_lengths)
>>> print(output)
[[3. 2. 1.]
 [5. 4. 6.]
 [9. 8. 7.]]
>>> x = Tensor(np.array([[1, 2, 3, 4], [5, 6, 7, 8]]), mindspore.float32)
>>> seq_lengths = Tensor(np.array([4, 4]))
>>> reverse_sequence = ops.ReverseSequence(seq_dim=1)
>>> output = reverse_sequence(x, seq_lengths)
>>> print(output)
[[4. 3. 2. 1.]
 [8. 7. 6. 5.]]