www.openml.org Open in urlscan Pro
131.155.11.11  Public Scan

Form analysis
0 forms found in the DOM

Text Content

You need to enable JavaScript to run this app.
OpenML
Search

Datasets
5.6k

Tasks
261.5k

Flows
17.9k

Runs
10.1M

Collections
198
Tasks
198

Runs
198

Benchmarks
28
Task Suites
28

Run Studies
28

Task Types
8

Measures
228
Data qualities
228

Eval. Measures
228

Eval. Procedures
228
Learn

Documentation

Blog

API's

Contribute

Meet up

About us

Terms & Citation




Minify

Dark

Search
Sign In
Sign Up

OpenML
A worldwide machine learning lab
Machine learning research should be easily accessible and reusable. OpenML is an
open platform for sharing datasets, algorithms, and experiments - to learn how
to learn better, together.
I shared a new data setI found a better model!OpenML
Datasets
Sign Up
to start tracking and sharing your own work.
OpenML is open and free to use.


AI-READY DATA

All datasets are uniformy formatted, have rich, consistent metadata, and can be
loaded directly into your favourite environments.


ML LIBRARY INTEGRATIONS

Pipelines and models can be shared directly from your favourite machine learning
libraries. No manual steps required.


A TREASURE TROVE OF ML RESULTS

Learn from millions of reproducible machine learning experiments on thousands of
datasets to make informed decisions.

FRICTIONLESS MACHINE LEARNING

Easily import and export datasets, pipelines, and experiments from your
favourite machine learning environments and libraries.



IN PYTHON (WITH SCIKIT-LEARN)


        from sklearn import ensemble
        from openml import tasks, runs

        clf = ensemble.RandomForestClassifier()
        task = tasks.get_task(3954)
        run = runs.run_model_on_task(clf, task)
        run.publish()
      

IN R (WITH MLR3)


        library(mlr3oml)
        library(mlr3)

        task = tsk("oml", task_id = 31)
        resampling = rsmp("oml", task_id = 31)

        resample(task, lrn("classif.rpart"), resampling)
      

Learn more about the OpenML APIs

ALL MACHINE LEARNING DATA, ORGANIZED


For every dataset, find which tasks (e.g. classification) need to be solved.

For every task, find all evaluation runs that people did, and how well their
models performed.

For every run, find model details, evaluations, and the exact algorithm
pipelines used.

For every flow (pipeline), find all the evaluation runs to see how well it
performed on different tasks.
Learn how OpenML works

REPRODUCIBLE MACHINE LEARNING


OpenML records exactly which datasets and library versions are used, so that
nothing gets lost.

For every experiment, the exact pipeline structure, architecture, and all
hyperparameter settings are automatically stored.

OpenML flows wrap around tool-specific implementations that can be serialized
and later deserialized to reproduce models and verify results.
Read stories
wraprebuildoriginal runreproduced run

LEARNING TO LEARN

Run systematic benchmarks, large-scale experiments, learn from previous
experiments, and automate machine learning itself.


BENCHMARKING SUITES

Easy-to-use, curated suites of machine learning tasks to standardize and improve
benchmarking.

DocumentationPaper


AUTOML BENCHMARK

An open, ongoing, and extensible benchmark framework for Automated Machine
Learning systems.

DocumentationPaper


AUTOML TOOLS

Several AutoML tools use OpenML to speed up the search for the best models,
includingautosklearnSageMaker AutoMLAzure AutoMLGAMA


LEARN TO TUNE

Learn from millions of experiments how to tune algorithms:parameter
importancedefault learningsymbolic defaults

Read more research based on OpenML

JOIN OPENML

Join a vibrant ecosystem of machine learning researchers and enthousiasts.


MACHINE LEARNING EXPERTS

Share your work, show how it's done, and track how often it is viewed and reused


DATA OWNERS

Share your data to challenge and collaborate with the machine learning community


ALGORITHM DEVELOPERS

Integrate your tools with OpenML to easily import and export data and
experiments


SOFTWARE ENGINEERS

OpenML is open source, get involved and make it even better and more useful

Learn how to contribute to OpenML

SUPPORT OPENML

We gratefully acknowledge the support from our sponsors and supporting
organizations.Become a sponsor



GET STARTED NOW

Sign Up
OpenML guides
Get Involved
About Us
 * 
   New dataset
 * 
   New task
 * 
   New collection


 * 
   Your profile
 * 
   Sign out