HRI (Repository of our Nao & Pepper robot research project)

Demo videos

How to execute the code

• Go to the folder chatgpt, execute the file hri_main.py.
• That should work otherwise, please activate the conda environment hri first and then execute this file.

My setup

• I am using windows 11 (WSL - Ubuntu 18)
• For gpu setup please install, nvidia cuda toolkit on windows
• Your distro will be able to access the gpu drivers
• You need to install USBIPD on windows
• Follow the steps from this website :
  • https://learn.microsoft.com/en-us/windows/wsl/connect-usb
• The audio drivers might not be present by default, need to install PulseAudio

Installation

To install this project, follow these steps:

• Install Miniconda (https://docs.conda.io/en/latest/miniconda.html)
• Go to https://www.aldebaran.com/en/support/nao-6/downloads-softwares
• In your home dir create a folder "naoqi" and then download the python SDK there and extract.
  • add "export PYTHONPATH=~/naoqi/pynaoqi-python2.7-2.8.6.23-linux64-20191127_152327/lib/python2.7/site-packages:$PYTHONPATH" to your .bashrc

Install PulseAudio, pavucontrol is for GUI

sudo apt update
sudo apt install pulseaudio -y
sudo apt install pavucontrol -y

Create Conda env

conda create --name hri
conda activate hri

Install PyTorch GPU

conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia

Install record audio libraries

conda install -c anaconda pyaudio

Install Voice Authentication module

pip install pyannote.audio

Install OpenAI for chatGPT integration

conda install -c conda-forge openai

Install Pyglet

conda install -c conda-forge pyglet

Install Whisper voice trascription model

pip install -U openai-whisper
pip install git+https://github.com/openai/whisper.git 
pip install --upgrade --no-deps --force-reinstall git+https://github.com/openai/whisper.git
sudo apt update && sudo apt install ffmpeg

This code needs API keys from OPENAI(for chatGPT) and huggingface.co(for pyannote)

• I have added these keys to the .bashrc file
• OpenAI link - https://platform.openai.com/account/api-keys

export OPENAI_API_KEY="you-key-please"

• HuggingFace link - https://huggingface.co/settings/tokens
• Also you have to agree to some T&C. Preferably run it 1st time on jupyter, you will get the link there itself.

export PYANNOTE_API_KEY="you-key-please"

You may need to create an ssh setup for GitHub

• Follow the comands below, you may discard the prompts
• copy and paste the id_rsa.pub contents to https://github.com/settings/keys

ssh-keygen -t rsa -b 4096 -C "email@domain.com"
eval $(ssh-agent -s)
ssh-add ~/.ssh/id_rsa
git config --global user.email "email@domain.com"
git config --global user.name "Jon Doe"