espnet2.asr.transducer.rnnt_multi_blank.rnnt_multi_blank.check_contiguous

source

espnet2.asr.transducer.rnnt_multi_blank.rnnt_multi_blank.check_contiguous(var, name)

Check if the given variable is contiguous in memory.

This function raises a ValueError if the input tensor is not contiguous. Contiguous tensors are stored in a single, contiguous block of memory, which can be more efficient for certain operations in PyTorch.

• Parameters:
  • var (torch.Tensor) – The tensor to check for contiguity.
  • name (str) – The name of the tensor being checked, used in the error message.
• Raises:ValueError – If the input tensor is not contiguous.

Examples

>>> import torch
>>> a = torch.tensor([[1, 2], [3, 4]])
>>> check_contiguous(a, "a")  # This will pass
>>> b = a.transpose(0, 1)
>>> check_contiguous(b, "b")  # This will raise a ValueError