Revamped version of the old archived NEURAL repository.
NEURAL2 is a torch-like collection of neural-network-related classes and structures used for doing simple machine learning tasks. It is modular and supports various different operations. Like VSC (the renderer I made previously), this is made entirely using software.
Modularity is based on PyTorch/Tensorflow. The computational algorithms are based on various resources obtained on the internet, some of which are credited in the runner files.
https://deeplearning.cs.cmu.edu/F21/
NULL
Only main.cpp for now.
Read through the header files for information on how neural networks learn and operate.
Feel free to push an issue, do a pull request, or simply contact me (NormalExisting on Discord, Bluesky, and Telegram).
Confusion matrix for conv1d unit test: [[4886, 1281], [1126, 12707]]