dbmdz BERT models
A collection of pre-trained BERT, DistilBERT, ELECTRA, GPT-2, and ConvBERT models for multiple languages, including German, Italian, Turkish, and historic texts.
What is dbmdz BERT models?
dbmdz/berts is a collection of transformer-based language models developed by the MDZ Digital Library team. It provides pre-trained models for various languages, including German, Italian, Turkish, French, Ukrainian, and historic languages, enabling natural language processing tasks without training from scratch. The project focuses on under-resourced languages and specialized domains like historic texts.
NLP researchers and developers working with specific languages like German, Italian, Turkish, French, Ukrainian, or historic languages such as Historic Dutch, Finnish, and Swedish. It is also aimed at those needing models for specialized domains like digitized newspapers and archival texts.
Developers choose this project for its specialized models for under-resourced languages and historic texts, which are not widely covered by generic multilingual models. The models are community-driven, optimized for specific languages, and fully compatible with the Hugging Face Transformers library for easy integration.
Overview
DBMDZ BERT, DistilBERT, ELECTRA, GPT-2 and ConvBERT models
Use Cases
Best For
- Performing NLP tasks on historic texts from sources like Europeana newspapers or the Delpher Corpus.
Related Projects
Found a gem we're missing?
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
