Is there currently a way to specify your own distance metric function for jobs with K-Means Clustering?

[CMagnoDFB]

I'm trying to perform some clustering tasks but my dataset is pretty big (very long biosignals). Even reducing the dimensionality, a main clustering job might take me longer than 3 days (I have limited compute capability). I've come across some variations for the dtw, one in particular called fastdtw, and wondered if its possible to swap the function.