There are some SO posts on calculating accuracy of a classification model in PyTorch, but I how do I calculate accuracy of an approximation model?
For example, for classifications, I can usually count, per-class, the amount of hits, divided by the total validation set size: correct_count / len(validation_set)