Revision history [back]

I was able to train with decreasing the numPos argument value. I set the with this formula numPos = 0.9 * number_of_positive_samples and 0.99 as a minHitRate. This time i was able to complete the training however my results are terrible which i will continue to work.