Build instructions
install apt dependencies
$sudo apt install unzip
install pyenv
$curl https://pyenv.run | bash
$exec "$SHELL" #restart shell
install build dependencies (https://github.com/pyenv/pyenv/wiki#suggested-build-environment)
$sudo apt update; sudo apt install build-essential libssl-dev zlib1g-dev \ libbz2-dev libreadline-dev libsqlite3-dev curl git \ libncursesw5-dev xz-utils tk-dev libxml2-dev libxmlsec1-dev libffi-dev liblzma-dev
$pyenv install 3.11.2
$pyenv global 3.11.2
setup venv
$python -m venv venv
source venv/bin/activate
get whisper-live
$wget https://github.com/collabora/WhisperLive/archive/refs/tags/v0.5.1.zip
$unzip v0.5.1.zip
install whisper-live
$sudo bash WhisperLive-0.5.1/scripts/setup.sh
$pip install --force-reinstall -v "whisper-live==0.5.1"
$pip install --force-reinstall -v "numpy<2>"
run the server and the client
$sh server.sh
$python client.py
installing cuDNN (for nvidia GPUs)
WhisperLive-0.5.1 uses ctranslate2 v4.5.0, which uses cuDNN 9. So, we must manually install the cudnn toolkit (V12) and then cudnn (v9)
Install nvidia drivers:
$sudo ubuntu-drivers install
$nvidia-smi #check drivers work
Installed cudnn drivers using apt:
$sudo apt install nvidia-cuda-toolkit
$nvcc --version #verify cuda install
Download cuDNN installer from nvidia website, and followed instructions for installing cuDNN 12
instructions at https://developer.nvidia.com/cudnn-downloads
$sudo apt-get -y install cudnn