import bisect

MOD = 10**9 + 7

class FenwickTree:
    def __init__(self, size):
        self.size = size
        self.tree = [0] * (size + 1)  # 1-based indexing

    def update(self, index, value):
        while index <= self.size:
            self.tree[index] = (self.tree[index] + value) % MOD
            index += index & -index

    def query(self, index):
        res = 0
        while index > 0:
            res = (res + self.tree[index]) % MOD
            index -= index & -index
        return res

def main():
    import sys
    input = sys.stdin.read().split()
    ptr = 0
    K = int(input[ptr])
    ptr += 1
    N = int(input[ptr])
    ptr += 1
    x = list(map(int, input[ptr:ptr+N]))
    ptr += N
    x.sort()  # Ensure x is sorted

    max_k = K
    fenwick = FenwickTree(max_k)
    dp = [0] * (max_k + 1)
    dp[0] = 1
    fenwick.update(0 + 1, dp[0])  # Fenwick is 1-based

    for k in range(1, max_k + 1):
        # Find the largest x_i <= k
        m = bisect.bisect_right(x, k) - 1
        if m < 0:
            dp_k = 0
        else:
            # Compute the sum of dp[k - x_i] for i=0 to m
            # Since x is sorted, k - x_i may be in a non-contiguous range
            # Thus, we need to query each x_i individually, which is O(m) time and not feasible
            # This approach is incorrect as it leads to O(N) per k
            # Hence, the code below is incorrect for the problem constraints
            sum_dp = 0
            for i in range(m + 1):
                xi = x[i]
                if xi > k:
                    break
                j = k - xi
                if j < 0:
                    continue
                sum_dp = (sum_dp + dp[j]) % MOD
            dp_k = sum_dp
        dp[k] = dp_k
        fenwick.update(k + 1, dp_k)

    print(dp[K] % MOD)

if __name__ == "__main__":
    main()