Skip to content

Linux Support

Jürgen Fleiß edited this page Dec 6, 2024 · 6 revisions

Installers

Deb file for beta 1.2.1 here.

Ubuntu (Debian) Manual Setup

Limitations

  • We do not currently provide a flatpack, snap or out of beta debian package. Installation and start via command line, but with the aTrain user interface.
  • Currently, only CPU transcription is supported reliably.
  • Installation of additional packages via apt from the Universe repository.
  • Currently tested on: Ubuntu 24.04 LTS

Setup

Update package lists:

sudo apt update

Install dependencies:

`sudo apt install ffmpeg python3 python3-pip python3-venv git build-essential libgl1-mesa-dev libcairo2 libcairo2-dev libgirepository1.0-dev libwebkit2gtk-4.0-dev -y

Depending on your Ubuntu version, you might need to install python from a third party repository:

sudo add-apt-repository ppa:deadsnakes/ppa

sudo apt install python3.11

Create a virtual environment in your home folder (or any other you prefer) and activate it.

cd ~

python3 -m venv atrain_venv

source atrain_venv/bin/activate

Now, install atrain from the github repository.

pip install git+https://github.com/JuergenFleiss/aTrain.git@v1.2.1 --extra-index-url https://download.pytorch.org/whl/cu121

Next, you need aTrain to download all the necessary models for transcription and speaker detection. This has only to be done once when installing it.

aTrain init

You are now ready to start aTrain. The following command should open a webrowser window with the user interface.

aTrain start

Start aTrain after installation

If you already installed aTrain, you can start it by:

cd ~

source atrain_venv/bin/activate

aTrain start

Clone this wiki locally