#include <cstdio>
#include <cstring>
#include <iostream>
#include <string>
#include <cmath>
#include <bitset>
#include <vector>
#include <map>
#include <set>
#include <queue>
#include <deque>
#include <algorithm>
#include <complex>
#include <unordered_map>
#include <unordered_set>
#include <random>
#include <cassert>
#include <fstream>
#define popcount __builtin_popcount
using namespace std;
typedef long long int ll;
typedef pair<int, int> Pii;
 
#define FOR(i,n) for(int i = 0; i < (n); i++)
#define sz(c) ((int)(c).size())
#define ten(x) ((int)1e##x)
 
template<class T> T extgcd(T a, T b, T& x, T& y) { for (T u = y = 1, v = x = 0; a;) { T q = b / a; swap(x -= q * u, u); swap(y -= q * v, v); swap(b -= q * a, a); } return b; }
template<class T> T mod_inv(T a, T m) { T x, y; extgcd(a, m, x, y); return (m + x % m) % m; }
ll mod_pow(ll a, ll n, ll mod) { ll ret = 1; ll p = a % mod; while (n) { if (n & 1) ret = ret * p % mod; p = p * p % mod; n >>= 1; } return ret; }
 
template<int mod, int primitive_root>
class NTT {
public:
	int get_mod() const { return mod; }
	void _ntt(vector<ll>& a, int sign) {
		const int n = sz(a);
		assert((n ^ (n&-n)) == 0); //n = 2^k
 
		const int g = 3; //g is primitive root of mod
		int h = (int)mod_pow(g, (mod - 1) / n, mod); // h^n = 1
		if (sign == -1) h = (int)mod_inv(h, mod); //h = h^-1 % mod
 
		//bit reverse
		int i = 0;
		for (int j = 1; j < n - 1; ++j) {
			for (int k = n >> 1; k >(i ^= k); k >>= 1);
			if (j < i) swap(a[i], a[j]);
		}
 
		for (int m = 1; m < n; m *= 2) {
			const int m2 = 2 * m;
			const ll base = mod_pow(h, n / m2, mod);
			ll w = 1;
			FOR(x, m) {
				for (int s = x; s < n; s += m2) {
					ll u = a[s];
					ll d = a[s + m] * w % mod;
					a[s] = u + d;
					if (a[s] >= mod) a[s] -= mod;
					a[s + m] = u - d;
					if (a[s + m] < 0) a[s + m] += mod;
				}
				w = w * base % mod;
			}
		}
 
		for (auto& x : a) if (x < 0) x += mod;
	}
	void ntt(vector<ll>& input) {
		_ntt(input, 1);
	}
	void intt(vector<ll>& input) {
		_ntt(input, -1);
		const int n_inv = mod_inv(sz(input), mod);
		for (auto& x : input) x = x * n_inv % mod;
	}
 
	// 畳み込み演算を行う
	vector<ll> convolution(const vector<ll>& a, const vector<ll>& b){
		int ntt_size = 1;
		while (ntt_size < sz(a) + sz(b)) ntt_size *= 2;
 
		vector<ll> _a = a, _b = b;
		_a.resize(ntt_size); _b.resize(ntt_size);
 
		ntt(_a);
		ntt(_b);
 
		FOR(i, ntt_size){
			(_a[i] *= _b[i]) %= mod;
		}
 
		intt(_a);
		return _a;
	}
};
 
typedef NTT<167772161, 3> NTT_1;
typedef NTT<469762049, 3> NTT_2;
typedef NTT<1224736769, 3> NTT_3;
 
// garnerのアルゴリズムを直書きしたversion，速い
vector<ll> fast_int32mod_convolution(vector<ll> a, vector<ll> b,int mod){
	for (auto& x : a) x %= mod;
	for (auto& x : b) x %= mod;
	
	NTT_1 ntt1; NTT_2 ntt2; NTT_3 ntt3;
	assert(ntt1.get_mod() < ntt2.get_mod() && ntt2.get_mod() < ntt3.get_mod());
	auto x = ntt1.convolution(a, b);
	auto y = ntt2.convolution(a, b);
	auto z = ntt3.convolution(a, b);
 
	// garnerのアルゴリズムを極力高速化した
	const ll m1 = ntt1.get_mod(), m2 = ntt2.get_mod(), m3 = ntt3.get_mod();
	const ll m1_inv_m2 = mod_inv<ll>(m1, m2);
	const ll m12_inv_m3 = mod_inv<ll>(m1 * m2, m3);
	const ll m12_mod = m1 * m2 % mod;
	vector<ll> ret(sz(x));
	FOR(i, sz(x)){
		ll v1 = (y[i] - x[i]) *  m1_inv_m2 % m2;
		if (v1 < 0) v1 += m2;
		ll v2 = (z[i] - (x[i] + m1 * v1) % m3) * m12_inv_m3 % m3;
		if (v2 < 0) v2 += m3;
		ll constants3 = (x[i] + m1 * v1 + m12_mod * v2) % mod;
		if (constants3 < 0) constants3 += mod;
		ret[i] = constants3;
	}
 
	return ret;
}

const ll MOD=1e9+7;
ll solve(vector<ll> p, int k){
    vector<ll> p0; p0.push_back(p[0]);
    vector<ll> ret(1, 1);
    int i=0;
    while((1<<i)<=k){
        for(int j=(1<<i); j<min(100001, 1<<(i+1)); j++){
            p0.push_back(p[j]);
        }
        i++;
        vector<ll> p1=fast_int32mod_convolution(ret, p0, MOD);
        p1.resize(1<<i);
        for(int j=0; j<p1.size(); j++){
            if(j==0) p1[0]=(2-p1[0]+MOD)%MOD;
            else p1[j]=(MOD-p1[j])%MOD;
        }
        ret=fast_int32mod_convolution(ret, p1, MOD);
        ret.resize(1<<i);
    }
    return ret[k];
}

int main()
{
    int k, n; cin>>k>>n;
    vector<ll> p(100001, 0);
    p[0]=1;
    for(int i=0; i<n; i++){
        int x; cin>>x;
        p[x]=MOD-1;
    }
    cout<<solve(p, k)<<endl;
    return 0;
}