The best model can be trained and evaluated through python main.py
latent_1_eval
and latent_2_eval
store the inference figures by latent dimensions of 1 and 2 respectively, where the names of figures stand for the value of latent vectors.
results_*
directories store the best models for a certain number of latent dimensions and the corresponding evaluation results throughout the whole training process.
Running evaluation.py
gives you inference figures.