This is a simple multi layer perceptron (mlp) or well knows as artificial neural network framework written in native python and only using numpy array library. this framework also provides 4 optimization algorithms. run MLP_main.py to run MLP model, run data.py to generate dummy data.
layer files are activation_relu.py, activation_softmax.py, layer_dense.py, and loss.py optimizer object classes are in optimizer.py file other files are experimental files so feel free to try write your own model!
- SGD
- RMS_prop
- Adagrad
- Adam(combination of RMSprop and adagrad)