42 Berlin AI - Quickstart

The goal of this project is to explore, test and understand the future of human-AI interactions, combining multiple sensory inputs and incorporating the latest technologies in generative AI and the visual arts. We make our developments openly available, to build a community of practice with fellows and students to shape and research the future of AI.

Prerequisites

• Unity v2021.3.1f1
• Tensorflow
• Python

Requirements

• API Key for OpenAI
• API Key for Google Cloud, allowing Speech-to-text and Text-to-speech
• API Key for Dialogflow
• API Key for OpenWeatherMap
• OpenCV for Unity (you can get a free demo)

Setup

Installations

1. Clone this repository

2. Create a new virtual environment from the server folder and activate it

   cd 42BearPlayground/42Berlin_AI_server;
   python -m venv venv;
   . venv/bin/activate;

3. Install the requirements, making sure you uncomment the correct lines according to your operating system

   pip install -r requirements.txt 

4. For MacOs only : install deepface

   pip install deepface --no-deps

Environment variables

1. For both folders (42Berlin_AI and 42Berlin_AI_server), make a copy of the example .env file and fill it with your secret keys

   cp .env.example .env

2. Add your Google cloud JSON file for Dialogflow as 42Berlin_AI_server/private_key.json

3. Add your Google cloud JSON file for Speech-to-text and Text-to-speech in 42Berlin_AI/Assets/Resources/private_key.json

4. Add a file containing some speech context (one sentence per line) as 42Berlin_AI/Assets/FrostweepGames/StreamingSpeechRecognition/speechContext.txt

Launch

1. In 42Berlin_AI_server/, run the server

   sh launch.sh

2. Open and run the 42Berlin_AI project on Unity

Useful links

• OpenAI website / How GPT-3 works
• Deepface repository / How faces recognition works using Deepface