Skip to content

App for monitoring conversation-share in real-time. Developed in the neuefische Data Science, Machine Learning & AI Bootcamp 2024 in Hamburg.

Notifications You must be signed in to change notification settings

jottemka/ds-capstone-speakequal

Repository files navigation

Speequal - Promoting Fair Conversations

alt text

Speequal is an app that gives users a real-time overview of the share each speaker has in a conversation. It is intended to help us make conversations more equitable and fair.

The app was developed as part of the neuefische Data Science, Machine Learning & AI Bootcamp 2024 in Hamburg by:

Notice, however, that the app is largely based on the diart framework:

Running The App

alt text

To run the app, navigate to the "streamlit" folder and execute the following command:

streamlit run app.py

You can then start monitoring a conversation via microphone or by feeding an audio signal into the app and then clicking the "Start streaming" button. To stop, simply click the "Stop streaming" button.

Fine-Tuning

There are a number of settings that can be fine-tuned for better performance:

  1. Detection Threshold for Active Speakers: high values mean that short utterances will not be detected as speech.
  2. Cluster-Update Threshold: low values mean that the cluster centroid is easier to shift.
  3. New-Speaker Threshold: lower values mean that more speaker labels will be generated.
  4. Maximum Number of Speakers: upper limit of speakers.

Environment Setup

This repo contains a requirements.txt file with a list of all the packages and dependencies you will need.

Before you can start with plotly in Jupyter Lab you have to install node.js (if you haven't done it before). Check Node version by run the following commands:

node -v

If you haven't installed it yet, begin at step_1. Otherwise, proceed to step_2.

macOS type the following commands :

Step_1: Update Homebrew and install Node by following commands:

brew update
brew install node

Step_2: Install the virtual environment and the required packages by following commands:

pyenv local 3.11.3
python -m venv .venv
source .venv/bin/activate
pip install --upgrade pip
pip install -r requirements.txt

WindowsOS type the following commands :

Step_1: Update Chocolatey and install Node by following commands:

choco upgrade chocolatey
choco install nodejs
  • Step_2: Install the virtual environment and the required packages by following commands.

For PowerShell CLI :

pyenv local 3.11.3
python -m venv .venv
.venv\Scripts\Activate.ps1
pip install --upgrade pip
pip install -r requirements.txt

For Git-Bash CLI :

pyenv local 3.11.3
python -m venv .venv
source .venv/Scripts/activate
pip install --upgrade pip
pip install -r requirements.txt

Note: If you encounter an error when trying to run pip install --upgrade pip, try using the following command:

python.exe -m pip install --upgrade pip

About

App for monitoring conversation-share in real-time. Developed in the neuefische Data Science, Machine Learning & AI Bootcamp 2024 in Hamburg.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages