Gravitational-wave parameter estimation with machine-learning generated surrogate waveforms