Instructions to use UHH-CI/GermanPolitical-Gelectra-base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use UHH-CI/GermanPolitical-Gelectra-base with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="UHH-CI/GermanPolitical-Gelectra-base")# Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("UHH-CI/GermanPolitical-Gelectra-base") model = AutoModelForSequenceClassification.from_pretrained("UHH-CI/GermanPolitical-Gelectra-base") - Notebooks
- Google Colab
- Kaggle
GermanPolitical-Gelectra-base
Released May 2022, we fine-tuned the german electra model pre-trained by german-nlp-group (orginal model can be found at 'german-nlp-group/electra-base-german-uncased') to classify political sentiment of popular german parties. Supported parties at the moment are "CDU/CSU", "SPD", "FDP", "Grüne", "Die Linke", "AfD".
Overview
Architecture: Gelectra-base (uncased)
Language: German
Datasets & Methodology
For more information about data and methodology visit: https://www.wahl-o-bot.de/Methodik
Production
The model can be used for free at https://www.wahl-o-bot.de
Authors
Feel free to contact us about questions or feedback. We are always happy to help.
- Maximilian Witte:
maximilian.witte@uni-hamburg.de - Jasper Schwenzow:
jasper.schwenzow@uni-hamburg.de - Mark Heitmann:
mark.heitmann@uni-hamburg.de - Matthias Assenmacher:
matthias@stat.uni-muenchen.de
The model is built in collaboration of the University of Hamburg and Ludwig Maximilians University Munich.
- Downloads last month
- 3
