Hello!

I reviewed examples of GraphCRF for OCR letter recognition in user guide.

Q1:I just wanted to check whether I got everything right - as I see the example for OCR letter recognition shows edge array as

array([[0, 1, 2, 3, 4, 5, 6, 7],
         [1, 2, 3, 4, 5, 6, 7, 8]]

The shape of array above is (2, n_edges), while the documentation state that we need array of shape (n_edges, 2).

Q2: Training of GraphCRF taking a lot of time (~7h with FrankoWolfeSVM, with others learners even more) on my data. Although I have near 300 data points. Is there any tips how I can speed up the process?

Thanks in advance,
Mariia.