Includes a series of jupyter notebooks to practice the following concepts:
- Build computation graph used to construct neural networks.
- Construct Value class as an atomic part of the computation graph.
- Build neurons, layers, and a complete neural network with multiple layers.
- Introduce multiple common diagnosis of a neural network model.
- Batch normalization and other techniques to resolve common issues in neural networks.
- Build character-level language models (inspired by karpathy/makemore).