In our current CV implementation, folds are evaluated sequentially. This is fine for small training sets, but for larger ones it becomes prohibitively expensive. The code needs to be modified to do the CV splitting as an initialization step, and then run the training of each fold in parallel. This needs to be implemented in cross_validation.py
In our current CV implementation, folds are evaluated sequentially. This is fine for small training sets, but for larger ones it becomes prohibitively expensive. The code needs to be modified to do the CV splitting as an initialization step, and then run the training of each fold in parallel. This needs to be implemented in
cross_validation.py