brainage04 / westpachackathon

This guide assumes that you are using Visual Studio Code.

Install Depedencies

Open this project in Visual Studio Code Open a terminal (`Ctrl + ``, should be the button to the left of the 1 key) Run the following commands:

1. (optional) python -m venv .venv (creates a virtual environment for storing dependencies, important on Linux to avoid overwriting older Python packages with newer ones as this can break your operating system/desktop environment)
2. (optional) Windows: source .venv/Scripts/activate | Linux: source .venv/bin/activate (activates the virtual environment so you can use and install dependencies)
3. python -m pip install "kivy[full]" kivy_examples SpeechRecognition tensorflow numpy IPython pvrecorder pydub PyAudio (this may take a while, it will download ~1GB worth of packages)

Reload the window (Ctrl + Shift + P > Developer: Reload Window). Done!

Test Project

To test the project, click on the front_end.py file and click the play button in the top right. Make sure that the virtual environment is activated (denoted by a (westpac_hackathon) prefix). If it is not, run the second command from the Install Dependencies section and try again.

Notes

Don't worry about warnings from the Tensorflow package. They are just warnings. The tensorflow package for Windows does not have GPU support, meaning that GPU support would require a Linux operating system, most likely through the Windows Subsystem for Linux (WSL) and building of the tensorflow package from scratch for your specific hardware (which I am not expecting anyone to put themselves through).

Todo

Modify machine learning model to use contrastive loss function Test own voice samples on modified model