There is bic/aic criterion with GMM in scikit-learn but I want to fit my data on GPU. I found GMM implemented in CuPy(cuda numpy wrapper) but it dosen't have bic/aic criterion.(https://github.com/cupy/cupy/blob/master/examples/gmm/gmm.py)
How to implement bic/aic criterion to that code? or Is there good libraries? Please help.
