Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/webis-de/small-text
Active Learning for Text Classification in Python
https://github.com/webis-de/small-text
active-learning deep-learning language-models looking-for-contributors machine-learning natural-language-processing nlp python pytorch small-language-models text-classification transformers
Last synced: 7 days ago
JSON representation
Active Learning for Text Classification in Python
- Host: GitHub
- URL: https://github.com/webis-de/small-text
- Owner: webis-de
- License: mit
- Created: 2021-05-24T08:06:41.000Z (over 3 years ago)
- Default Branch: main
- Last Pushed: 2024-11-24T18:39:14.000Z (18 days ago)
- Last Synced: 2024-11-28T20:04:22.308Z (14 days ago)
- Topics: active-learning, deep-learning, language-models, looking-for-contributors, machine-learning, natural-language-processing, nlp, python, pytorch, small-language-models, text-classification, transformers
- Language: Python
- Homepage: https://small-text.readthedocs.io/
- Size: 3.22 MB
- Stars: 593
- Watchers: 25
- Forks: 65
- Open Issues: 15
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
Awesome Lists containing this project
- awesome-active-learning - Small-text - learn, pytorch) | Christopher Schröder | Text classification | (3.3 AL in AI Fields - 人工智能背景中的主动学习 / **Tutorials - 教程**)
README
[](https://pypi.org/project/small-text/)
[](https://anaconda.org/conda-forge/small-text)
[](https://codecov.io/gh/webis-de/small-text)
[](https://small-text.readthedocs.io/en/v2.0.0.dev1/)
[](CONTRIBUTING.md)
[](LICENSE)
[](https://zenodo.org/records/14212697)
> Active Learning for Text Classification in Python.
[Installation](#installation) | [Quick Start](#quick-start) | [Contribution](CONTRIBUTING.md) | [Changelog][changelog] | [**Docs**][documentation_main]
Small-Text provides state-of-the-art **Active Learning** for Text Classification.
Several pre-implemented Query Strategies, Initialization Strategies, and Stopping Critera are provided,
which can be easily mixed and matched to build active learning experiments or applications.
## What is Active Learning?
[Active Learning](https://small-text.readthedocs.io/en/latest/active_learning.html) allows you to efficiently label training data for supervised learning in a scenario where you have little to no labeled data.
## Features
- Provides unified interfaces for Active Learning so that you can
easily mix and match query strategies with classifiers provided by [sklearn](https://scikit-learn.org/), [Pytorch](https://pytorch.org/), or [transformers](https://github.com/huggingface/transformers).
- Supports GPU-based [Pytorch](https://pytorch.org/) models and integrates [transformers](https://github.com/huggingface/transformers)
so that you can use state-of-the-art Text Classification models for Active Learning.
- GPU is supported but not required. In case of a CPU-only use case,
a lightweight installation only requires a minimal set of dependencies.
- Multiple scientifically evaluated components are pre-implemented and ready to use (Query Strategies, Initialization Strategies, and Stopping Criteria).
---
## News
**Version 2.0.0 dev1** ([v2.0.0.dev1][changelog_2.0.0dev1]) - November 24th, 2024
- This a development release with the most changes so far. You can consider this an alpha release, which does not guarantee you stable interfaces yet,
but is otherwise ready to use.
- Version 2.0.0 offers cleaned up interfaces, new query strategies, improved classifiers, and new functionality such as vector indices. See the [changelog][changelog_2.0.0dev1] for a full list of changes.
**Version 1.4.1** ([v1.4.1][changelog_1.4.1]) - August 18th, 2024
- Bugfix release.
**Version 1.4.0** ([v1.4.0][changelog_1.4.0]) - June 9th, 2024
- New query strategy: [AnchorSubsampling](https://small-text.readthedocs.io/en/v1.3.3/components/query_strategies.html#small_text.query_strategies.subsampling.AnchorSubsampling) (aka [AnchorAL](https://arxiv.org/abs/2404.05623)).
Special thanks to [Pietro Lesci](https://github.com/pietrolesci) for the correspondence and code review.
**Paper published at EACL 2023 🎉**
- The [paper][paper_published] introducing small-text has been accepted at [EACL 2023](https://2023.eacl.org/). Meet us at the conference in May!
- Update: the paper was awarded [EACL Best System Demonstration](https://aclanthology.org/2023.eacl-demo.11/). Thank you, for your support!
[For a complete list of changes, see the change log.][changelog]
---
## Installation
Small-Text can be easily installed via pip:
```bash
pip install small-text
```
The command results in a [slim installation][documentation_install] with only the necessary dependencies.
For a full installation via pip, you just need to include the `transformers` extra requirement:
```bash
pip install small-text[transformers]
```
The library requires Python 3.8 or newer. For using the GPU, CUDA 10.1 or newer is required.
More information regarding the installation can be found in the
[documentation][documentation_install].
## Quick Start
For a quick start, see the provided examples for [binary classification](examples/examplecode/binary_classification.py),
[pytorch multi-class classification](examples/examplecode/pytorch_multiclass_classification.py), and
[transformer-based multi-class classification](examples/examplecode/transformers_multiclass_classification.py),
or check out the notebooks.
### Notebooks
| # | Notebook | |
| --- |----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| 1 | [Intro: Active Learning for Text Classification with Small-Text](https://github.com/webis-de/small-text/blob/v2.0.0.dev1/examples/notebooks/01-active-learning-for-text-classification-with-small-text-intro.ipynb) | [](https://colab.research.google.com/github/webis-de/small-text/blob/v2.0.0.dev1/examples/notebooks/01-active-learning-for-text-classification-with-small-text-intro.ipynb) |
| 2 | [Using Stopping Criteria for Active Learning](https://github.com/webis-de/small-text/blob/v2.0.0.dev1/examples/notebooks/02-active-learning-with-stopping-criteria.ipynb) | [](https://colab.research.google.com/github/webis-de/small-text/blob/v2.0.0.dev1/examples/notebooks/02-active-learning-with-stopping-criteria.ipynb) |
| 3 | [Active Learning using SetFit](https://github.com/webis-de/small-text/blob/v2.0.0.dev1/examples/notebooks/03-active-learning-with-setfit.ipynb) | [](https://colab.research.google.com/github/webis-de/small-text/blob/v2.0.0.dev1/examples/notebooks/03-active-learning-with-setfit.ipynb) |
| 4 | [Using SetFit's Zero Shot Capabilities for Cold Start Initialization](https://github.com/webis-de/small-text/blob/v2.0.0.dev1/examples/notebooks/04-zero-shot-cold-start.ipynb) | [](https://colab.research.google.com/github/webis-de/small-text/blob/v2.0.0.dev1/examples/notebooks/04-zero-shot-cold-start.ipynb) |
### Showcase
- [Tutorial: 👂 Active learning for text classification with small-text][argilla_al_tutorial] (Use small-text conveniently from the [argilla][argilla] UI.)
A full list of showcases can be found [in the docs][documentation_showcase].
🎀 **Would you like to share your use case?** Regardless if it is a paper, an experiment, a practical application, a thesis, a dataset, or other, let us know and we will add you to the [showcase section][documentation_showcase] or even here.
## Documentation
Read the latest documentation [here][documentation_main]. Noteworthy pages include:
- [Overview of Query Strategies][documentation_query_strategies]
- [Reproducibility Notes][documentation_reproducibility_notes]
---
## Scope of Features
Extension of Table 1 in the EACL 2023 paper.
Name
Active Learning
Query Strategies
Stopping Criteria
small-text v1.3.0
14
5
small-text v2.0.0
19
5
We use the numbers only to show to tremendous progress that small-text has made over time.
There many features and improvements that are not reflected in these numbers.
## Alternatives
[modAL](https://github.com/modAL-python/modAL), [ALiPy](https://github.com/NUAA-AL/ALiPy), [libact](https://github.com/ntucllab/libact), [ALToolbox](https://github.com/AIRI-Institute/al_toolbox)
---
## Contribution
Contributions are welcome. Details can be found in [CONTRIBUTING.md](CONTRIBUTING.md).
## Acknowledgments
This software was created by Christopher Schröder ([@chschroeder](https://github.com/chschroeder)) at Leipzig University's [NLP group](http://asv.informatik.uni-leipzig.de/)
which is a part of the [Webis](https://webis.de/) research network.
The encompassing project was funded by the Development Bank of Saxony (SAB) under project number 100335729.
## Citation
Small-Text has been introduced in detail in the EACL23 System Demonstration Paper ["Small-Text: Active Learning for Text Classification in Python"](https://aclanthology.org/2023.eacl-demo.11/) which can be cited as follows:
```
@inproceedings{schroeder2023small-text,
title = "Small-Text: Active Learning for Text Classification in Python",
author = {Schr{\"o}der, Christopher and M{\"u}ller, Lydia and Niekler, Andreas and Potthast, Martin},
booktitle = "Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations",
month = may,
year = "2023",
address = "Dubrovnik, Croatia",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.eacl-demo.11",
pages = "84--95"
}
```
## License
[MIT License](LICENSE)
[documentation_main]: https://small-text.readthedocs.io/en/v2.0.0.dev1/
[documentation_install]: https://small-text.readthedocs.io/en/v2.0.0.dev1/install.html
[documentation_query_strategies]: https://small-text.readthedocs.io/en/v2.0.0.dev1/components/query_strategies.html
[documentation_showcase]: https://small-text.readthedocs.io/en/v2.0.0.dev1/showcase.html
[documentation_reproducibility_notes]: https://small-text.readthedocs.io/en/v2.0.0.dev1/reproducibility_notes.html
[changelog]: https://small-text.readthedocs.io/en/latest/changelog.html
[changelog_1.4.0]: https://small-text.readthedocs.io/en/latest/changelog.html#version-1-4-0-2024-06-09
[changelog_1.4.1]: https://small-text.readthedocs.io/en/latest/changelog.html#version-1-4-1-2024-08-18
[changelog_2.0.0dev1]: https://small-text.readthedocs.io/en/latest/changelog.html#version-2-0-0-dev1-2024-11-24
[argilla]: https://github.com/argilla-io/argilla
[argilla_al_tutorial]: https://docs.argilla.io/en/latest/tutorials/notebooks/training-textclassification-smalltext-activelearning.html
[paper_published]: https://aclanthology.org/2023.eacl-demo.11v2.pdf