A Feedforward Neural Net used to classify the Iris Dataset. You can find the Iris Species Dataset here. This Neural Net was implemented for educational purposes in order to understand Backpropagation.
You can experiment with the model's learning rate, number of epochs, number of neurons per layer and the type of loss function (Mean Square or Categorical Cross Entropy). You can also change the train/test split ratio by changing the TRAIN_SIZE and TEST_SIZE constants at irisdata/iris_load.h.
Run the Neural Net with:
$ sh run_nn.sh