mlr-org/mlr3

mlr3: Machine Learning in R - next generation

What it solves

mlr3 is a modern, object-oriented framework for machine learning in R. It provides a standardized way to handle the building blocks of machine learning—such as tasks, learners, and resampling strategies—to avoid the feature creep and maintenance difficulties associated with older frameworks.

How it works

The project uses an object-oriented design based on R6 classes to manage state changes and reference semantics. It leverages data.table for high-performance data frame computations and checkmate for strict type safety and argument checking. The core package focuses strictly on computation, while specialized functionality (like visualization or specific learners) is offloaded to extension packages.

Who it’s for

Data scientists and researchers using R who need a professional, extensible, and efficient machine learning pipeline for classification, regression, and survival analysis.

Highlights

  • Object-Oriented Design: Uses R6 for a clean, queryable architecture.
  • High Performance: Integrated with data.table for fast data handling.
  • Extensible Ecosystem: A wide array of extension packages for learners, tuning, and pipelines.
  • Type Safety: Implements defensive programming to ensure predictable return types and stable user input.
  • Parallelization: Supports parallel backends via the future abstraction for resampling and benchmarking.

Related

  • Project
    isl-org/Open3D-ML

    An extension of Open3D that provides tools, pretrained models, and pipelines for 3D machine learning tasks like semantic segmentation and object detection.

  • Project
    PrimeIntellect-ai/prime-rl

    A framework for large-scale asynchronous reinforcement learning designed to train frontier models with 1T+ parameters on up to 1000+ GPUs.

  • Project
    openml/automlbenchmark

    A framework for evaluating and comparing open-source AutoML systems using standardized environments and curated datasets to ensure reproducible research.

  • Project
    dotnet/machinelearning

    A cross-platform open-source machine learning framework for .NET developers to build, train, and deploy custom models without needing Python or R.

  • Project
    hiyouga/EasyR1

    An efficient and scalable reinforcement learning training framework that extends support to vision language models and multiple RL algorithms like GRPO.