// clang-format off
#include <bits/stdc++.h>
using namespace std;
using ll = long long;
using ld = long double;
template<class T> using V = vector<T>;
using VI = V<int>;
using VL = V<ll>;
using VS = V<string>;
template<class T> using PQ = priority_queue<T, V<T>, greater<T>>;
using G = V<VI>;
template<class T> using WG = V<V<pair<int, T>>>;
#define overload4(_1,_2,_3,_4,name,...) name
#define overload3(_1,_2,_3,name,...) name
#define rep1(n) for(ll _=0;_<(n);++_)
#define rep2(i,n) for(ll i=0;i<(n);++i)
#define rep3(i,a,b) for(ll i=(a);i<(b);++i)
#define rep4(i,a,b,c) for(ll i=(a);i<(b);i+=(c))
#define rep(...) overload4(__VA_ARGS__,rep4,rep3,rep2,rep1)(__VA_ARGS__)
#define rrep1(n) for(ll i=n-1;i>=0;--i)
#define rrep2(i,n) for(ll i=n-1;i>=0;--i)
#define rrep3(i,a,b) for(ll i=b-1;i>=a;--i)
#define rrep(...) overload3(__VA_ARGS__,rrep3,rrep2,rrep1)(__VA_ARGS__)
#define all(a) a.begin(),a.end()
#define rall(a) a.rbegin(),a.rend()
#define inside(h,w,y,x) (unsigned(y)<h&&unsigned(x)<w)
#ifdef _DEBUG
#define line cout << "-----------------------------\n"
#define debug(...) print(__VA_ARGS__)
#else
#define line
#define debug
#endif
constexpr ll INF = 1000000000;
constexpr ll LLINF = 1LL << 61;
constexpr ll mod = 1000000007;
constexpr ll MOD = 998244353;
constexpr ld eps = 1e-10;
constexpr int dy[]{ -1,0,1,0 }, dx[]{ 0,1,0,-1 };
template<class T> inline bool chmax(T& a, const T& b) { if (a < b) { a = b; return true; }return false; }
template<class T> inline bool chmin(T& a, const T& b) { if (a > b) { a = b; return true; }return false; }
inline void init() { cin.tie(nullptr); cout.tie(nullptr); ios::sync_with_stdio(false); cout << fixed << setprecision(15); }
template<class T> inline istream& operator>>(istream& is, V<T>& v) { for (auto& a : v)is >> a; return is; }
template<class T, class U> inline istream& operator>>(istream& is, pair<T, U>& p) { is >> p.first >> p.second; return is; }
template<class T> inline V<T> vec(size_t a) { return V<T>(a); }
template<class T> inline V<T> defvec(T def, size_t a) { return V<T>(a, def); }
template<class T, class... Ts> inline auto vec(size_t a, Ts... ts) { return V<decltype(vec<T>(ts...))>(a, vec<T>(ts...)); }
template<class T, class... Ts> inline auto defvec(T def, size_t a, Ts... ts) { return V<decltype(defvec<T>(def, ts...))>(a, defvec<T>(def, ts...)); }
template<class T> inline void print(const T& a) { cout << a << "\n"; }
template<class T, class... Ts> inline void print(const T& a, const Ts&... ts) { cout << a << " "; print(ts...); }
template<class T> inline void print(const V<T>& v) { for (int i = 0; i < v.size(); ++i)cout << v[i] << (i == v.size() - 1 ? "\n" : " "); }
template<class T> inline void print(const V<V<T>>& v) { for (auto& a : v)print(a); }
template<class T> inline constexpr const T cum(const V<T>& a, int l, int r) { return 0 <= l && l <= r && r < a.size() ? a[r] - (l == 0 ? 0 : a[l - 1]) : 0; }//[l,r]
template<class T> inline constexpr const auto min(const T& v) { return *min_element(all(v)); }
template<class T> inline constexpr const auto max(const T& v) { return *max_element(all(v)); }
template<class T> inline V<T>& operator++(V<T>& v) { for (T& a : v)++a; return v; }
template<class T> inline V<T>& operator--(V<T>& v) { for (T& a : v)--a; return v; }
// clang-format on

#include <atcoder/all>
using namespace atcoder;
using mint = atcoder::modint998244353;

class combination {
public:
    int n;
    ll c_mod;
    VL fac, finv, inv;

    combination(int n, ll c_mod) : n(n), c_mod(c_mod), fac(n), finv(n), inv(n) {
        fac[0] = fac[1] = 1;
        finv[0] = finv[1] = 1;
        inv[1] = 1;
        for (int i = 2; i < n; ++i) {
            fac[i] = fac[i - 1] * i % c_mod;
            inv[i] = c_mod - inv[c_mod % i] * (c_mod / i) % c_mod;
            finv[i] = finv[i - 1] * inv[i] % c_mod;
        }
    }
    ll operator()(int n, int k) {
        if (n < k) return 0;
        if (n < 0 || k < 0) return 0;
        return fac[n] * (finv[k] * finv[n - k] % c_mod) % c_mod;
    }
};

int main() {
    init();

    int m, n;
    cin >> m >> n;

    VI c(m);
    rep(i, n) {
        int a;
        cin >> a;
        --a;
        ++c[a];
    }
    VI cc(n + 1);
    for (int x : c) ++cc[x];
    V<mint> g(n + 1, 1);
    combination com(n + 10, MOD);
    rep(j, n + 1) if (cc[j] > 0) {
        int x = j, s = cc[j];
        rep(i, 1, n + 1) {
            if (i - x >= 0) g[i] *= mint(com.fac[i] * com.finv[i - x]).pow(s);
            else g[i] = 0;
        }
    }

    mint ans = 0;
    rep(l, 1, n + 1) {
        mint tmp = 0;
        rep(i, 1, l + 1) {
            mint res = (mint)com.finv[i] * com.finv[l - i] * g[i];
            tmp += res * ((l - i) & 1 ? -1 : 1);
        }
        ans += tmp * com.fac[l];
    }
    print(ans.val());

    return 0;
}