Python implementation of simple example of Neural Network. No high level libary is needed. Use numpy, pandas and matplotlib.

The project was created using Anaconda. You can recreate an environment using environment.yml file (see: https://conda.io/docs/using/envs.html#use-environment-from-file).

The code from the repository demonstrates how to build a simple neural network (forward pass, backward pas). Data set of bike sharing is also part of repository so to show a training and perofmance evaluation actions.

Feel free to contact me in case of questions.

References:

1. http://ufldl.stanford.edu/tutorial/supervised/SoftmaxRegression/
2. http://cs231n.github.io/neural-networks-case-study/#grad

Please join me if you wnat to learn-while-doing. Just drop me a message. What topics to cover in the future here:

1. type of activation functions
2. backpropagations with more details
3. weights update (how it happends during a training)
4. Using validations in training
5. How do I know I arrived? (is it training already done?)
6. Testing, that is final performance evaluation
7. adding more referneces if necessary