PyTorch DTW C++ extension

Dynamic time warping in native PyTorch, with CPU and CUDA backends.

pip install torchdtw

This package requires PyTorch 2.10 or later. It is developed using the PyTorch 2.10 Stable ABI, and compiled with instructions for CUDA cards from Volta to Blackwell. This was originally made for fastabx, but it can be used in other projects. Only the exact DTW is implemented, there is no plan to add variants.

Usage

This package provides three functions:

DTW

def dtw(distances: torch.Tensor) -> torch.Tensor

Compute the DTW cost of the given distances 2D tensor.

Arguments:

• distances: A 2D tensor of shape (n, m) representing the pairwise distances between two sequences.

Returns:

A scalar tensor with the cost.

DTW path

def dtw_path(distances: torch.Tensor) -> torch.Tensor

Compute the DTW path of the given distances 2D tensor.

No CUDA variant or batched implementation are provided for now.

Arguments:

• distances: A 2D tensor of shape (n, m) representing the pairwise distances between two sequences.

Returns:

A 2D tensor of shape (*, 2) with the path indices.

Batched DTW

def dtw_batch(distances: torch.Tensor, sx: torch.Tensor, sy: torch.Tensor, *,
              symmetric: bool) -> torch.Tensor

Compute the batched DTW cost on the distances 4D tensor.

Arguments:

• distances: A 4D tensor of shape (n1, n2, s1, s2) representing the pairwise distances between two batches of sequences.
• sx: A 1D tensor of shape (n1,) representing the lengths of the sequences in the first batch.
• sy: A 1D tensor of shape (n2,) representing the lengths of the sequences in the second batch.
• symmetric: Whether or not the DTW is symmetric (i.e., the two batches are the same).

Returns:

A 2D tensor of shape (n1, n2) with the costs.

Benchmark

Check this folder for comparisons against reference implementations.

Citation

Please cite the fastabx paper if you use this package in your work:

@misc{fastabx,
  title={fastabx: A library for efficient computation of ABX discriminability},
  author={Maxime Poli and Emmanuel Chemla and Emmanuel Dupoux},
  year={2025},
  eprint={2505.02692},
  archivePrefix={arXiv},
  primaryClass={cs.CL},
  url={https://arxiv.org/abs/2505.02692},
}