Literature · OpenMed
OpenMed OrganismDetect
A compact biomedical transformer model for detecting species and organism mentions in scientific literature.
What OpenMed OrganismDetect is
OpenMed OrganismDetect (OpenMed-NER-OrganismDetect-TinyMed-82M) is a token-classification model developed by OpenMed for named entity recognition (NER) in biomedical and scientific texts. It extracts organism mentions, including scientific binomials, common names, genera, strains, and informal references.
Built to assist researchers in navigating vast biological and healthcare corpora, the model automates the extraction of biological taxa from published abstracts and clinical texts. This facilitates downstream analyses such as biodiversity informatics, systematic biology, literature mining for drug discovery, and the construction of biomedical knowledge graphs.
How it works
The model is based on the distilroberta-base architecture, an 82-million parameter transformer fine-tuned for token classification. It processes tokenised biomedical text and assigns BIO-format entity tags (B-SPECIES and I-SPECIES) to identify the boundaries of organism mentions.
Training was conducted using the Hugging Face Transformers framework with the AdamW optimizer and learning rate scheduling, validated via cross-validation on a held-out test split. The model was fine-tuned on the SPECIES800 corpus, a curated benchmark containing 800 biomedical abstracts manually annotated for taxonomic mentions across diverse biological kingdoms.
Key facts
- Base architecture
- distilroberta-base
- Parameter count
- 82M
- Task
- Named entity recognition (token classification)
- Entity labels
- B-SPECIES, I-SPECIES
- Training corpus
- SPECIES800 (800 annotated abstracts)
- Language
- English
- Performance
- F1: 0.72, Precision: 0.72, Recall: 0.73, Accuracy: 0.95
What you can use it for
- Biodiversity and taxonomic literature mining
- Extracting scientific binomials and common organism names from abstracts to support biodiversity database curation and conservation research.
- Biomedical knowledge graph generation
- Parsing organism mentions from scientific literature to link biological model systems and pathogens to experimental outcomes, diseases, or targets.
- Clinical and pharmacovigilance text analysis
- Identifying microbial strains, pathogens, and biological model references across medical records, healthcare documents, and adverse event reports.
Run it here
No installation and no GPU of your own: fill in the inputs and it runs on our servers. Results stay in your browser unless you sign in.
Runs on On-demand GPU
Runs on our servers with on-demand compute. A first run needs time to load the model; active capacity can be reused and scales down when idle.Run this model
OpenMed OrganismDetect
Strengths and limits
Good at
- Identifies both formal scientific nomenclature and informal common organism names.
- Operates at an 82M parameter size, offering lower computational overhead compared to larger biomedical transformers.
- Achieves 0.95 overall token accuracy on the SPECIES800 benchmark.
Watch out for
- Delivers moderate recognition performance with an F1 score of 0.72, precision of 0.72, and recall of 0.73 on its evaluation benchmark.
- Restricted to identifying single-class organism boundaries (SPECIES) and does not differentiate between taxonomic ranks or entity types at the label level.
Inputs and outputs
You provide
Text
Entities are labelled token by token.
You get back
Labels with confidence scores, highest first.
Licence
Check the licence before commercial use · See model card
The See model card licence carries terms worth reading before you depend on it.
Questions
Can I use OpenMed OrganismDetect commercially?
Check the licence before commercial use. The See model card licence carries terms worth reading before you depend on it.
How do I run OpenMed OrganismDetect without installing anything?
Use the form under “Run it here” on this page: fill in the inputs and press Run model. It runs on our servers on CPU only, typically in 12s, and a run costs about 28 credits. Visitors without an account get 1,000 credits a month.
What entity tags does this model output?
The model outputs standard BIO tags for a single entity class: B-SPECIES for the beginning of an organism mention and I-SPECIES for continuation tokens.
How does OrganismDetect differ from typical species recognition models?
OrganismDetect is trained to capture a broader range of organism references than pure species extractors, including taxonomic genera, specific strains, and informal organism mentions.
What dataset was used to fine-tune the model?
It was fine-tuned on the SPECIES800 dataset, an annotated corpus comprising 800 biomedical abstracts covering diverse taxonomic groups.
How to cite it
Panahi, M. (2025). OpenMed NER: Open-Source, Domain-Adapted State-of-the-Art Transformers for Biomedical NER Across 12 Public Datasets. arXiv:2508.01630.
Written from OpenMed OrganismDetect's own documentation (model card). Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.