Modular Inferencing Utilities

[alexander-huang-commits]

Summary:
This is a modular way to inference with one script across different experiment types and initializations (FBP, SIRT, and whatever initializer you would like to add). Modular in the sense that when you add new metrics or neural network models, you just update the list at the top of the inferencing script.

Changes Made:
-Added a statistics file that prompts for models and metrics to compare them with using ANOVA
-Added a generalized inferencing script with better instructions in the top of the file (saves specified metrics to csv and/or images of best/worst performance)
-Added a script to save npy images of all models available in ALL_MODELS folder
-Added a parameter counting script to see size of NN
-Added CSV file tree for each experiment type

How to use:
There are readme's in each folder explaining the file structure and usage of each script. But in general, when a new model is made:

1. Use MODELNAME_test.py to create script that will save data to csv
2. Use general_stats.py to compare with other models