Skip to content
@possible-worlds-research

Possible Worlds Research

We research algorithms for energy-efficient AI.

Possible World Research is dedicated to researching and developing tiny AI models in the area of Natural Language Processing. Why tiny? Because we believe that tomorrow's technology should not pollute our beautiful planet. Also because small models can be used by everybody, irrespective of their financial resources. All our code is tested on consumer hardware: typically, home laptops and single board computers. No GPUs involved.

Developing tiny models means creating more intelligent algorithms. For any given task, a model that achieves high performance with fewer resources is arguably more intelligent. We believe that simulating human faculties in machines requires a good understanding of different scientific fields, including linguistics, psychology, and of course computer science. We welcome contributions that demonstrate some curiosity about the human mind.

Pinned Loading

  1. wikinlp Public

    A package to download and preprocess a Wikipedia dump, in any language.

    Python 9 1

  2. pretrained-tokenizers Public

    Multingual pretrained Sentencepiece tokenizers, obtained through wikinlp.

    1 1

  3. pearsnlp Public

    A package to do NLP work for the PeARS search engine.

    Python

Repositories

Showing 10 of 10 repositories
  • vgnlp Public

    A package to download and process the Visual Genome dataset

    Python 1 AGPL-3.0 0 0 0 Updated Mar 24, 2025
  • pretrained-tokenizers Public

    Multingual pretrained Sentencepiece tokenizers, obtained through wikinlp.

    1 1 0 0 Updated Mar 4, 2025
  • wikinlp Public

    A package to download and preprocess a Wikipedia dump, in any language.

    Python 9 AGPL-3.0 1 0 0 Updated Feb 21, 2025
  • pearsnlp Public

    A package to do NLP work for the PeARS search engine.

    Python 0 AGPL-3.0 0 0 0 Updated Feb 19, 2025
  • valencechat Public
    Python 1 0 0 0 Updated Nov 7, 2024
  • AI-buster Public

    Some scripts to demonstrate the instability of Transformers / LLMs.

    Python 0 1 0 0 Updated Apr 17, 2024
  • .github Public

    The README for the organization.

    0 0 0 0 Updated Jan 21, 2024
  • common-crawl-processor Public

    Code to process Common Crawl dumps

    Python 0 GPL-3.0 1 1 0 Updated Jan 21, 2024
  • SCSS 0 MIT 0 0 0 Updated Jan 10, 2024
  • ontoloader Public
    Python 0 AGPL-3.0 0 0 0 Updated Dec 26, 2023

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading…

Most used topics

Loading…