Provably Fast Convergence of Independent Natural Policy Gradient for Markov Potential Games