| | --- |
| | language: |
| | - de |
| | metrics: |
| | - f1 |
| | pipeline_tag: text-classification |
| | license: gpl-3.0 |
| | --- |
| | |
| | # GePart: German Party Classification Model |
| |
|
| | **GePart** is a German party classification model, being capable of predicting the most likely party to a given text. This repository is part of a student project at the [Baden-Württemberg Cooperative State University](https://www.dhbw.de/startseite) during the 5th and 6th semester of our bachelor's degree in computer science. |
| |
|
| | For training the classification model, we used multiple sources (e.g., speeches, tweets, and party programs) with word embeddings like [FastText](https://fasttext.cc/) and transformer models like [BERT](https://arxiv.org/abs/1810.04805). |
| |
|
| | ## Hosted Model |
| |
|
| | The hosted model is based on a German version of the DistilBERT Model. |
| |
|
| | ## License |
| |
|
| | This project is licensed under [GNU GPL V3](https://github.com/felixhoffmnn/gepart/blob/main/LICENSE). |
