Why my training model(K-Fold) return Nan Value?

Question

So I\'ve trained a model using 10 K-fold in Spyder, at K-fold(iterations 1-3) training model. Everything was fine with at least 99% accuracy, 98-99% val_a