Currently in engineering school, I'm trying to learn by myself how NN works and how libraries like Pytorch or Keras are working.
That's why I'm trying to implement a NN C# library from scratch, just for educational purposes.
VSCode
.NET 5.0
In order to get a local copy, follow these steps
- Dotnet ⩾ 5.0
and/or
- Visual Studio 2019
- Clone the repository :
git clone https://github.com/andreamiele/NNNET.git
cd NNNET
- Open the solution NNNET.csproj with VS or VSCode with .NET core and solution explorer extensions.
- Add Tensor Class with operations related
- Add Cost function
- Add Sigmoid, ReLu and Leaky ReLu functions
- Create a Layer class
- Create metric classes
- Add Optimizer (like SGB, the base optimizer or also Adam)
- Add Loss
- Add Autograd
- Add model
- Add Demos : xor, circles classification, MNIST, ...
- (?)