A repo that allows users to ask questions to Pepper, which are interpretted by Whisper, answered by ChatGPT, and the reply is spoken by Pepper
Download the repository and install the python dependencies (whisper, chatgpt, sounddevice, Paramiko).
Note that to set up ChatGPT you must export your API key.
Edit gpt.py to add the IP address, username and password of your Pepper robot.
Run the included choregraphe project on your pepper robot. You may wish to turn off "Autonomous Life" in Choregraphe first, and then wake pepper up using the "sun" button.
Next execute the gpt.py script in the terminal.
Once the whisper model has downloaded, wait for the "listening..." message then ask your question followed by silence.
Note that if you use a free API key, you will be limited to three questions per minute.
You can see the included example video here https://github.com/SkylarWiderski/Pepper_Chat_GPT/blob/main/examples/VID_20230707_164528.mp4