eval_models_layouts throws an error for sparse csr_matrix

[sciwithrach]

Line 162 in pipes.py should change from:

data=X.todense()

to

data = X.toarray()

Thanks so much for this package, it's amazing!