Should I chose k = 3 in this case of KNN classifier?

Question

I have data with the dimension of (2055, 95). I split it into train data: (1640, 95) and validation data: (415, 95).

I build a KNN classifier but don\'t know which k