https://github.com/berenslab/text-embed-augm
Companion repository to our González-Márquez et al. (2025) preprint "Cropping outperforms dropout as an augmentation strategy for training self-supervised text embeddings" (arXiv)
Science Score: 36.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
○CITATION.cff file
-
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
○DOI references
-
✓Academic publication links
Links to: arxiv.org -
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (5.8%) to scientific vocabulary
Repository
Companion repository to our González-Márquez et al. (2025) preprint "Cropping outperforms dropout as an augmentation strategy for training self-supervised text embeddings" (arXiv)
Basic Info
- Host: GitHub
- Owner: berenslab
- License: mit
- Language: Jupyter Notebook
- Default Branch: main
- Homepage: https://arxiv.org/abs/2508.03453
- Size: 121 MB
Statistics
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
- Releases: 0
Metadata Files
README.md
Cropping outperforms dropout as an augmentation strategy for training self-supervised text embeddings
Rita González-Márquez, Philipp Berens & Dmitry Kobak
In this repository you can find the code associated to the paper "Cropping outperforms dropout as an augmentation strategy for training self-supervised text embeddings".
How to use this repository
The notebooks in scripts/ contain the code to reproduce all the experiments and analyses performed in the paper.
The notebook 08-rgm-figures.ipynb contains the code to generate the final figures included in the paper. All figures generated with the notebooks will be stored in the results/figures/updated_dataset/final_figures/last_version folder.
Installation
This project depends on Python ($\geq$ 3.7). The project script can be installed via pip install . in the project root, i.e.:
git clone https://github.com/berenslab/text-embed-augm
cd text-embed-augm
pip install -e .
Owner
- Name: Berens Lab @ University of Tübingen
- Login: berenslab
- Kind: organization
- Email: philipp.berens@uni-tuebingen.de
- Location: Tübingen, Germany
- Website: https://hertie.ai/data-science
- Repositories: 60
- Profile: https://github.com/berenslab
Department of Data Science at the Hertie Institute for AI in Brain Health, University of Tübingen
GitHub Events
Total
- Push event: 3
Last Year
- Push event: 3