A from-scratch implementation of a simple neural network in C++ using MSE loss and stochastic gradient descent.
Everything is contained within a single C++ file which contains heavily commented code explaining how the network code is formulated.
To compile and run in Windows,
g++ Network.cpp
.\aProject roadmap
- Neuron/Network Class
- Backpropagation
- Data class
- Custom loss function
- Custom optimisers