Which modification to the space-optimized DP code below correctly counts the number of such subsets?

Suppose the problem changes: now you can use each element in the array unlimited times to form subsets summing to K (i.e., unlimited repetitions allowed). Which modification to the space-optimized DP code below correctly counts the number of such subsets?

def count_subsets_unlimited(arr, K):
    dp = [0] * (K + 1)
    dp[0] = 1
    for num in arr:
        for j in range(???):
            dp[j] += dp[j - num]
    return dp[K]

Choose the correct range for the inner loop to handle unlimited usage of elements.