The Verilog code describes a multi-layer perceptron with pre-trained XOR logic weights and biases in the testbench.
Architecture of the MLP is as follow:
- 2 Inputs
- 1 Hidden Layer: 3 2-input neurons with ReLU activation function
- 1 Output Layer: 1 output neuron with ReLU activation function
P.S. code is not optimised yet.