AI-on-Demand - obtain and share AI resources with ease

• quick instantiation of AI models from string IDs across the ecosystem
• cross-platform metadata catalogue for papers, projects, datasets, models
• marketplace features for sharing models and assets (under development)

A web interface for assets is available through the AIoD catalogue and the Metadata Catalogue Editor (under construction)

Overview	AIoD Homepage · Tutorials · Release Notes
Open Source
Community
CI/CD
Code

Installation

Install from PyPI via:

$ pip install aiondemand

Quick start

AI model instantiation

Also see the AI model instantiation tutorial

aiod.get(id: str) is the key entry point:

• id is a unique string describing a model
• return is the python object, or an exception informing the user of required dependencies

from aiod import get

clf = get("RandomForestClassifier(n_estimators=42)")
clf

directly constructs the random forest from scikit-learn

RandomForestClassifier(n_estimators=42)

Works across the ecosystem!

from aiod import get

clf = get("XGBClassifier()")
clf

... constructs the XGBClassifier, or raises an informative error message to install xgboost in the environment (if not available)

XGBClassifier()

Complex pipelines and composites are supported out-of-the-box:

spec = """
pipe = Pipeline(steps=[
    ("imputer", SimpleImputer(strategy="mean")),
    ("scaler", StandardScaler()),
    ("classifier", RandomForestClassifier(n_estimators=100))])
cv = KFold(n_splits=5, shuffle=True, random_state=42)

return GridSearchCV(
    estimator=pipe,
    param_grid=[{
        "classifier__max_depth": [5, 10],
        "classifier__min_samples_split": [2, 5],
    },
    ],
    cv=cv,
    )
"""

get(spec)

• no more hassle with import statements!
• share your AI model specs as easy-to-handle strings!

Indexed libraries

The following libraries are indexed - the scope is expanding, roadmap contributions are welcome!

scikit-learn tabular estimators

• scikit-learn
• catboost
• feature-engine
• lightgbm
• imbalanced-learn
• mlxtend
• scikit-lego
• xgboost

Probabilistic supervised learning and survival modelling

• skpro

Time series, forecasting

• sktime

AI asset catalogue search

Also see the AI metadata catalogue tutorial

You can directly access endpoints through the Python API, for example to browse datasets:

import aiod

aiod.datasets.get_list()

And results will be returned as a Pandas dataframe (though the data_format may be used to get JSON instead):

      platform platform_resource_identifier                    name       date_published                                            same_as  is_accessible_for_free  ...  relevant_link  relevant_resource relevant_to research_area scientific_domain identifier
0  huggingface       acronym_identification  acronym_identification  2022-03-02T23:29:22  https://huggingface.co/datasets/acronym_identi...                    True  ...             []                 []          []            []                []          1
...
9  huggingface              allegro_reviews         allegro_reviews  2022-03-02T23:29:22    https://huggingface.co/datasets/allegro_reviews                    True  ...             []                 []          []            []                []         10

[10 rows x 30 columns]

You can even query the elastic search endpoints:

aiod.publications.search(query="Robotics")

      platform platform_resource_identifier                                               name date_published                                            same_as is_accessible_for_free  ... relevant_resource relevant_to      research_area  scientific_domain  type  identifier
0  robotics4eu                         1803  Responsible Robotics & non-tech barriers t...           None  https://www.robotics4eu.eu/publications/respon...                   None  ...                []          []  [other materials]  [other materials]  None           4

[1 rows x 36 columns]

Contributing

Interested in contributing? Check out the contributing guidelines, then start with the Developer Setup Guide to set up your local development environment. Check out the complete documentation here: https://aiondemand.github.io/aiondemand/developer_setup By contributing to this project, you agree to abide by our Code of Conduct.

Credits

The aiondemand package is being developed with funding from EU’s Horizon Europe research and innovation program under grant agreement No. 101070000 (AI4EUROPE). Not all contributors are affiliated with this funding.

cookiecutter and the py-pkgs-cookiecutter template were used to create the repository structure.