Transfer Learning of High-Fidelity Opacity Spectra in Autoencoders and Surrogate Models