ambn / m,n>=1
The strings in this language are {ab, aab, abb, aabb, aaab, abbb ... }. The language for this would be :
S -> aS / aB
B -> bB / b
we can make a NFA for this language so it is regular.
anbn / n>=1
The strings in this language will be { ab, aabb, aaabbb ...}. For every string in this language we have to check whether the number of a's are equal to the number of b's which would require some memory to store the number of a's for checking with b's which cannot be done by a NFA so this is not regular.