metadata
license: cc-by-sa-4.0
Flair-abbr-roberta-pubmed-plos-unfiltered
This is a stacked model of embeddings from roberta-large, HunFlair pubmed models and character-level language models trained on PLOS, fine-tuning on the PLODv2 unfiltered dataset. It is released with our LREC-COLING 2024 publication Using character-level models for efficient abbreviation and long-form detection. It achieves the following results on the test set:
Results on abbreviations:
- Precision: 0.8977
- Recall: 0.9351
- F1: 0.9160
Results on long forms:
- Precision: 0.8726
- Recall: 0.9260
- F1: 0.8985