Installer grt avec Homebrew

Liens du projet

• Page d'accueil
• Dépôt
• Page du gestionnaire de paquets

Commandes et alias

• grt-lin-reg-tool
• grt-log-reg-tool
• grt-merge-tool
• grt-mlp-tool
• grt-rf-tool
• grt-softmax-tool
• grt-split-tool
• grt-svm-tool
• grt-test-tool

Historique du projet

GRT grew out of Nicholas Gillian and Joseph A. Paradiso's work at the MIT Media Lab Responsive Environments Group and was described in a 2014 Journal of Machine Learning Research paper. The paper presents it as a cross-platform C++ library for real-time gesture recognition, feature extraction, automatic gesture spotting, classification, regression, and clustering.

The GitHub repository was created in January 2014 and the public release stream on GitHub includes version 0.1.0 in February 2016, 0.2.0 in August 2016, and 0.2.4 in December 2016. The repository README describes both a C++ API and GUI, plus .grt and CSV data/model files.

Historique d'adoption

GRT found a niche in creative coding, interactive music, sensor interfaces, and human-computer interaction rather than broad systems administration. Its README lists extensions and users such as ofxGrt for openFrameworks, ml-lib for Max and Pure Data, and ESP, showing adoption through research and media-arts tooling.

Modes d'utilisation

The Homebrew package exposes small command-line tools for common models and data preparation tasks, while the library itself is normally embedded in C++ projects. Users build pipelines from preprocessing, feature extraction, and classifier modules, train models from labeled sensor data, and run predictions in real time.

Pourquoi les passionnés de paquets s'y intéressent

GRT is interesting in package-manager culture because it packages academic interactive-machine-learning work as a reusable native library plus command-line helpers. It sits at the boundary between research code, creative-coding ecosystems, and reproducible Unix installation.

Chronologie

• 2014: GRT described in the Journal of Machine Learning Research.
• 2014: GitHub repository created for nickgillian/grt.
• 2016: GRT Version 0.1.0 published on GitHub releases.
• 2016: GRT Version 0.2.0 published.
• 2016: GRT Version 0.2.4 published.

Related projects

• ofxGrt integrates GRT with openFrameworks.
• ml-lib uses GRT concepts in Max and Pure Data contexts.
• ESP is listed by the GRT project as an application using GRT as a backend machine-learning system.

Sources

• https://api.github.com/repos/nickgillian/grt
• https://api.github.com/repos/nickgillian/grt/releases?per_page=30
• https://github.com/nickgillian/grt
• https://nickgillian.com/grt/
• https://www.jmlr.org/papers/volume15/gillian14a/gillian14a.pdf