N.N.N. V.0.1-t

This version of N.N.N. replaces the previous version (N.N.N. V.0.1b). This program is only meant for training the model; it doesn’t use PyTorch and instead has its own custom structure.
Model structure: 350 -> 360 -> 370 -> 380 -> 370 -> 360 -> 350

There are two zip files provided:
N.N.N.-V-0.1-t.zip (Contains the actual code)
N.N.N.-V-0.1-t(exe).zip (Contains the .exe version for usage)
Only one needs to be downloaded for it to work.

Needed files
TestData (folder): This folder is where you should put the data files (in .json) so the program can access them. There are a few datasets available for usage by HamirinTK here.
Settings.json
Cached_data_files (folder): This folder will contain the .json data files that are now .txt after they were converted to a number format.
If you are using the .exe version, make sure these files are in the same folder with the executable.

Disclaimer
This uses numerical differentiation and is expected to be slow; an N.N.N.-T version will be released in the future for (hopefully) faster speeds.

For more information, you may see the documentation here.