Practical overview
When it fits
Use it when you want a Docker-hosted, browser-accessible workspace that brings notebook and coding tools together with common machine-learning libraries. For multi-user workspace management, the project points to ML Hub.
Who it is for
- Machine-learning developers who want notebooks, code editors and common data-science libraries in one environment.
- Data-science practitioners who want to run a browser-accessible workspace on their own machine or server.
- Teams that want to manage multiple user workspaces through the separately identified ML Hub project.
Documented uses
- Develop machine-learning and data-science projects in Jupyter, JupyterLab or browser-based VS Code.
- Use pre-installed libraries including TensorFlow, PyTorch, Keras and scikit-learn.
- Work with notebook-oriented Git integration and monitor training and hardware through TensorBoard and Netdata.
- Access a Linux desktop or connect to the environment through documented SSH and VNC options.
Requirements & limits
Docker is required. The README documents a single-instance Docker image and local browser access through port 8080.
- The documented deployment depends on Docker rather than a standalone desktop installer.
- The project describes ML Workspace as single-user; its README recommends ML Hub for multi-user setups.
- A productive-use example mounts the working directory into
/workspaceand configures Jupyter token authentication. - The repository lists an Apache-2.0 license.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Start a local ML Workspace instance
Launch one workspace container and open its browser interface locally.
Before you start
- Docker installed on the machine.
Steps
Run the command below.
docker run -p 8080:8080 mltooling/ml-workspace:0.13.2Open
http://localhost:8080after the container starts.
Expected outcome
A single ML Workspace instance is exposed on port 8080 for local browser access.
Guide limitations
- The command starts a basic single instance; the README gives a separate, more configured command for productive use.
- For a workspace running on another machine or a different exposed port, use that machine's IP or DNS name and the exposed port.
Run a persistent single-instance workspace
Start a background instance with a mounted project folder, token authentication and restart behavior.
Before you start
- Docker installed on the machine.
- A current working directory to mount as the workspace folder.
- A token value to set for
AUTHENTICATE_VIA_JUPYTER.
Steps
Run the documented Docker command with port mapping
8080:8080and nameml-workspace.Mount the current directory at
/workspaceand setAUTHENTICATE_VIA_JUPYTERto the chosen token.Include
--shm-size 512mand--restart always, as in the documented example.Open the workspace through the mapped port.
Expected outcome
A background container with the current directory available at `/workspace`, token-based Jupyter authentication, 512 MB shared memory and an always-restart policy.
Guide limitations
- The source presents this as a recommended baseline for productive single-instance use, not a complete deployment or security review.
- For remote hosting, access it using the host's IP or DNS name and exposed port.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 14 October 2026
- Original website title
- ml-tooling/ml-workspace · Official homepage title · retained in its source language
Related options
ML Hub — The project recommends ML Hub for multi-user setups; it is based on JupyterHub and manages and proxies workspace instances.