First project of Udacity Deep Learning Nanodegree. Trained a deep neural network from scratch to predict the number of bikes shared for a bike sharing company. Implemented forward, backward propagation and adjusted the hyperparameters to get the best predictions.
The file 'neural_network_from_scratch.ipynb' is the original notebook of the project.
The file 'my_answers.py' is the file where the NeuralNetworks model class is implemented. The forward and backward propagation functions are written and the functions to train the model, update the weights using gradient descent, and run the forward pass in the model are also written. The results from this file are called in the 'neural_network_from_scratch.ipynb' where the result of predictions on test data are plotted and thus, the performance of the model is seen.