Create README.md

README.md

+# Transformer tuning pipeline
+
+## Overview
+This Python script was used to train Transformer models on a corpus annotated for emotion. The annotation scheme is structured around four dimensions established in the evaluative model of emotion developed by Loeser ([2019](#1)). See Noblet ([2025](#2)) for a description of the annotation process and some analysis.
+
+## Installation
+
+To get started, clone this repository and install the dependencies:
+
+```bash
+git clone https://github.com/heutzi/TRANSFORMER_TUNING.git
+pip install -r requirements.txt
+```
+
+## Usage
+
+### Data Preparation
+Ensure your dataset is properly formatted following the examples provided in the folders *trains* and *test*s, in *corpora*.
+
+### Training
+Run the training script with the following command (replacing Alibaba-NLP/gte-multilingual-base with any compatible model):
+
+```bash
+python script_classifier_piped.py Alibaba-NLP/gte-multilingual-base
+```
+
+## References
+<a id="1">Loeser, F. (2019).</a>
+Modélisation probabiliste de l’influence des émotions sur l’acceptabilité des inno-vations [Université Grenoble Alpes (ComUE)]. https://www.theses.fr/2019GREAH003
+
+<a id="1">Noblet J. (2025).</a>
+Le bruit dans la mesure de la composante cognitive de l’émotion pour l’évaluation de l’acceptabilité des innovations. Corpus, 26. Publisher : Bases, corpus et langage - UMR 6039, DOI : <a href="https://journals.openedition.org/corpus/9267" target="_blank">10.4000/1364u.</a>
+
+## License
+This project is licensed under the MIT License.
+
+## Author
+Jonas Noblet