Text this: Independent Learning of Nash Equilibria in Partially Observable Markov Potential Games with Decoupled Dynamics