MLJ.jl

Spec sheet

Purpose

MLJ.jl is a Julia package for general, composable machine learning at scale.Read 12 Oct 2026 · juliaml.ai

Audience

The project welcomes contributors who can demonstrate prior experience in Julia and machine learning.Read 11 Oct 2026 · juliaml.ai

Free plan

YesRead 23 Sep 2026 · github.com

Integrations

MLJ's ecosystem includes OpenML.jl integration for the OpenML data science exchange platform.Read 12 Oct 2026 · juliaml.ai

Security

The official About page describes the license, project history, and governance, but does not state security or compliance certifications.Read 11 Oct 2026 · juliaml.ai

Support

The documentation directs users to Julia Discourse, GitHub issues, or the #mlj channel in Julia Slack for questions and reports.Read 12 Oct 2026 · juliaai.github.io

License

MLJ is open source and released under the MIT license.Read 12 Oct 2026 · juliaml.ai

Acceleration

Supported workflows can use CPU threading or multiprocessing; serial CPU processing is the default.Read 12 Oct 2026 · juliaai.github.io

Additional packages

Users generally need to add model-providing packages to access model-specific code, and MLJ's interactive load command can prompt for a provider.Read 12 Oct 2026 · juliaai.github.io

Bayesian analysis

YesRead 23 Sep 2026 · github.com

Composition

MLJ supports pipelines, stacks, and learning networks for composing models.Read 12 Oct 2026 · juliaai.github.io

Core functions

MLJ provides a common interface and meta-algorithms for selecting, tuning, evaluating, composing, and comparing machine learning models.Read 12 Oct 2026 · juliaai.github.io

Data import formats

CSV (CSV.jl), DataFrames/Tables.jl tables, RDatasets, OpenMLRead 23 Sep 2026 · github.com

Data support

MLJ works with data supported by the Tables.jl interface when Tables.istable(X) is true.Read 12 Oct 2026 · juliaai.github.io

Evaluation and tuning

MLJ provides syntax to tune and evaluate composite models and supports probabilistic predictions.Read 12 Oct 2026 · juliaai.github.io

Installation

MLJ is installed with Julia's package manager using Pkg.add("MLJ").Read 12 Oct 2026 · juliaai.github.io

Learning networks

Learning networks can train multiple models together and can be exported as standalone model types.Read 11 Oct 2026 · juliaml.ai

Model coverage

The documentation says MLJ supports over 180 machine learning models written in Julia and other languages, including many scikit-learn models.Read 12 Oct 2026 · juliaai.github.io

Model registry

Its model registry contains metadata for models that users can search and load on demand.Read 11 Oct 2026 · juliaml.ai

Model support

The MLJ ecosystem includes support for neural network models built with Flux.jl.Read 11 Oct 2026 · juliaml.ai

Pipelines

Pipelines can combine model components, and their parameters can be tuned together.Read 11 Oct 2026 · juliaml.ai

Project history

The MLJ project began at the Alan Turing Institute in 2019 as a Tools, Practices and Systems project.Read 12 Oct 2026 · juliaml.ai

Regression models

YesRead 23 Sep 2026 · github.com

Scripting support

YesRead 23 Sep 2026 · github.com

Tracking integration

MLJ workflows can log evaluation and tuning results to MLflow through MLJFlow.jl, which must be added separately; the documentation warns that this integration is under reassessment due to MLflow REST API issues.Read 12 Oct 2026 · juliaai.github.io

Tuning

MLJ includes hyperparameter optimization strategies such as Grid and RandomSearch.Read 11 Oct 2026 · juliaml.ai

Workflow

MLJ is an entry point for choosing, loading, composing, evaluating, and tuning machine learning models.Read 11 Oct 2026 · juliaml.ai

Where Specifiction places it

• Statistical Analysis Software#72 of 1067.4