I'm making an implementation of the softmax regression and I'm struggling to understand the nature behind the problem of increasing value of Cross-Entropy: $H(y_i, p_i)=-\sum_{i=1}^C y_i log(p_i)$, along with an increasing accuracy:
This is utmost confusing to me, because there's no class imbalance:
I am not entirely sure whether the sample size of $N = 112$ (at the very least I cannot prove it) is at fault. I will appreciate any help on the matter. Thank you in advance.
