Requirements to Install and upgrade Label Studio

System requirements

You can install Label Studio on a Linux, Windows, or MacOSX machine running Python 3.6 or later.

Port requirements

Label Studio expects port 8080 to be open by default. To use a different port, specify it when starting Label Studio. See Start Label Studio.

Server requirements

Allocate disk space according to the amount of data you plan to label. As a benchmark, 1 million labeling tasks take up approximately 2.3GB on disk when using the SQLite database. 50GB of disk space is recommended for production instances.

Use a minimum of 8GB RAM, but 16GB RAM is recommended. for example, t3.large or t3.xlarge on Amazon AWS.

For more on using Label Studio at scale and labeling performance, see Start Label Studio.

Software requirements

PostgreSQL version 11.5 or SQLite version 3.35 or higher.

Install prerequisite

Install Label Studio in a clean Python environment. Heartex highly recommends using a virtual environment (venv or conda) to reduce the likelihood of package conflicts or missing packages.

Install with pip

To install Label Studio with pip and a virtual environment, you need Python version 3.8 or later. Run the following:

python3 -m venv env
source env/bin/activate
python -m pip install label-studio

To install Label Studio with pip, you need Python version 3.8 or later. Run the following:

pip install label-studio

After you install Label Studio, start the server with the following command:

label-studio

The default web browser opens automatically at http://localhost:8080 with Label Studio. See start Label Studio for more options when starting Label Studio.

Install with Docker

Label Studio is also available as a Docker container. Make sure you have Docker installed on your machine.

Install with Docker on *nix

To install and start Label Studio at http://localhost:8080, storing all labeling data in ./my_project directory, run the following:

docker run -it -p 8080:8080 -v $(pwd)/mydata:/label-studio/data heartexlabs/label-studio:latest

Install with Docker on Windows

Or for Windows, you have to modify the volumes paths set by -v option.

Override the default Docker install

You can override the default Docker install by appending new arguments.

In Windows Command Line (cmd):

docker run -it -p 8080:8080 -v %cd%/mydata:/label-studio/data heartexlabs/label-studio:latest label-studio --log-level DEBUG

In PowerShell:

docker run -it -p 8080:8080 -v ${PWD}/mydata:/label-studio/data heartexlabs/label-studio:latest label-studio --log-level DEBUG

Build a local image with Docker

If you want to build a local image, run:

docker build -t heartexlabs/label-studio:latest .

Run with Docker Compose

Use Docker Compose to serve Label Studio at http://localhost:8080. You must use Docker Compose version 1.25.0 or higher.

Start Label Studio:

docker-compose up -d

This starts Label Studio with a PostgreSQL database backend. You can also use a PostgreSQL database without Docker Compose. See Set up database storage.

Install Label Studio without internet access

Download label-studio docker image (host with internet access and docker):

docker pull heartexlabs/label-studio:latest

Export it as a tar archive:

docker save heartexlabs/label-studio:latest | gzip > label_studio_latest.tar.gz

Transfer it to another VM:

scp label_studio_latest.tar.gz <ANOTHER_HOST>:/tmp

SSH into and import the archive:

docker image import /tmp/label_studio_latest.tar.gz

Follow steps from Install and Upgrade to run LS.

Install on Ubuntu

To install Label Studio on Ubuntu and run it in a virtual environment, run the following command:

python3 -m venv env
source env/bin/activate
sudo apt install python3.9-dev
python -m pip install label-studio

Install from source

If you want to use nightly builds or extend the functionality, consider downloading the source code using Git and running Label Studio locally:

git clone https://github.com/HumanSignal/label-studio.git
cd label-studio
# Install all package dependencies
pip install -e .
# Run database migrations
python label_studio/manage.py migrate
# Start the server in development mode at http://localhost:8080
python label_studio/manage.py runserver

Install with Anaconda

conda create --name label-studio
conda activate label-studio
pip install label-studio

Upgrade Label Studio

To upgrade to the latest version of Label Studio, reinstall or upgrade using pip.

pip install --upgrade label-studio

Migration scripts run when you upgrade to version 1.0.0 from version 0.9.1 or earlier.

To make sure an existing project gets migrated, when you start Label Studio, run the following command:

label-studio start path/to/old/project

The most important change to be aware of is changes to rename “completions” to “annotations”. See the updated JSON format for completed tasks.

If you customized the Label Studio Frontend, see the Frontend reference guide for required updates to maintain compatibility with version 1.0.0.