Why not using checkpoint['g_best_state'] in evaluate_model.py?

agrimgupta92 / sgan

Code for "Social GAN: Socially Acceptable Trajectories with Generative Adversarial Networks", Gupta et al, CVPR 2018

MIT License

813 stars 261 forks source link

Why not using checkpoint['g_best_state'] in evaluate_model.py? #42

Closed mingbocui closed 4 years ago

mingbocui commented 5 years ago

checkpoint['g_best_state'] is corresponding to the model weights with the lowest validation error, why you just use checkpoint['g_state'], which is the final g_state saved in the training. The final g_state is not guaranteed to be the best one. Hope someone could give an explanation. Thanks!