Code in this repository was used for teaching purposes.
- load given data examples
- how to augment data to leverage small dataset and enhance accuracy
- how to find an adequat network architecture (neural architecture search)
- define the network architecture using a state-of-the-art neural network framework (Tensorflow)
- get familiar with Google Colab and free to use compute power, also using hardware acceleration on GPU and TPU
- save and load trained model for later use
- integrate trained model into a simple python application
- integrate trained model into a web application
- deployment of the web application to a web server
For best performance use Google Chrome web browser!
Note: Also runs on smartphones, performance may be not as good as on a computer.