distilroberta-base-climate-f

Model Card for distilroberta-base-climate-f

Model Description

This is the ClimateBERT language model based on the FULL-SELECT sample selection strategy.

Note: We generally recommend choosing this language model over those based on the other sample selection strategies (unless you have good reasons not to). This is also the only language model we will update from time to time.

Using the DistilRoBERTa model as starting point, the ClimateBERT Language Model is additionally pre-trained on a text corpus comprising climate-related research paper abstracts, corporate and general news and reports from companies. The underlying methodology can be found in our language model research paper.

Update September 2, 2022: Now additionally pre-trained on an even larger text corpus, comprising >2M paragraphs. If you are looking for the language model before the update (i.e. for reproducibility), just use an older commit like 6be4fbd.

Climate performance model card

Citation Information

@inproceedings{wkbl2022climatebert,
    title={{ClimateBERT: A Pretrained Language Model for Climate-Related Text}},
    author={Webersinke, Nicolas and Kraus, Mathias and Bingler, Julia and Leippold, Markus},
    booktitle={Proceedings of AAAI 2022 Fall Symposium: The Role of AI in Responding to Climate Challenges},
    year={2022},
    doi={https://doi.org/10.48550/arXiv.2212.13631},
}