In both methods below,
A is the input vector (row or column)
B is the output vector of consecutive 1 counts, row vector.
Method 1 assumes the vector only contains 1s and 0s
A = [0 1 1 1 0 0 1 0 1 1 0 1];
B = diff(find([0,A(:).',0]==0))-1;
B(B==0) = []
Method 2 counts consecutive 1s, other elements can take any numeric value.
A = [0 1 1 1 0 0 1 0 1 1 0 1];
dA = diff([0,A(:).',0]==1);
B = find(dA==-1)-find(dA==1)
