Mathieu Blondel

`reference/dtw.py` is a NumPy implementation, a PyTorch implementation is not available yet. `theta` should be `m x n` distance matrix where `m` and `n` are the lengths of the two...

The message you're showing seems to be a warning, not an error. Did you run `python setup.py install`? The steps are ``` cd soft-dtw make cython python setup.py build sudo...

If you want an implementation that doesn't require any compilation / install, you can also try this [project](https://github.com/google-research/soft-dtw-divergences).

@arthurmensch is also interested in this. I guess you two could collaborate!

Closing since a PyTorch implementation is now available in https://github.com/arthurmensch/didyprog thanks to @arthurmensch.

@shaform I agree. Does anyone want to tackle it? I won't be able to do it myself, since I don't have the use case myself. Here are the things to...

Awesome! Thanks for sharing.

Hi! As you can see [here](https://github.com/mblondel/soft-dtw/blob/master/examples/plot_chainer_MLP.py#L67), I'm looping over the batch in the example code. So `SoftDTWLoss` is between two time series, not between two batches of time series. I...

`SoftDTWLoss` expects time series of shape `(length, n_dimensions)`, so if your time series are one dimensional and have shape `(length,)`, you indeed need a reshape.

I think the right way to do it would be to add a band constraint, as done in [Fast Global Alignment Kernels](http://www.marcocuturi.net/Papers/cuturi11fast.pdf) by @marcocuturi. This would allow to only compute...