Literature · OpenMed
OpenMed SpeciesDetect
A 109M-parameter transformer for detecting species and organism names in biomedical literature and clinical texts.
What OpenMed SpeciesDetect is
OpenMed SpeciesDetect is a token-classification model developed by OpenMed for named entity recognition (NER) of organisms and taxonomic references. It is trained to identify both scientific and common species names in biological literature, research articles, and clinical documents.
Accurate identification of species mentions is essential for biological text mining, ecological studies, comparative genomics, and biodiversity informatics. The model automates the extraction of organism references, helping researchers systematically index mentions across large corpuses of biomedical text.
The model outputs standard BIO tags representing the beginning and continuation of species entities (B-SPECIES and I-SPECIES). It achieves an F1 score of 0.78, with a precision of 0.72 and a recall of 0.84 on its benchmark evaluation.
How it works
The model is built on an ElectraMed 109M backbone (with an e5-base-v2 base architecture) and is configured for token classification. It accepts tokenised biomedical text and produces token-level predictions with confidence scores, categorising organism names using B-SPECIES and I-SPECIES tags.
Training was conducted using the Hugging Face Transformers framework on the LINNAEUS corpus, a specialised dataset where species mentions are annotated and normalised to NCBI Taxonomy identifiers. Optimisation was performed using the AdamW optimiser with learning rate scheduling, evaluated via cross-validation on a held-out test set.
Key facts
- Model size
- 109M parameters
- Base architecture
- e5-base-v2 / ElectraMed 109M
- Task
- Token classification (Named entity recognition)
- Dataset
- LINNAEUS corpus
- Output labels
- B-SPECIES, I-SPECIES
- Precision
- 0.72
- Recall
- 0.84
- F1 score
- 0.78
- Accuracy
- 0.98
- Language
- English
What you can use it for
- Taxonomic literature mining
- Extracting scientific and common organism names from scientific papers to support biodiversity informatics and taxonomic indexing.
- Comparative genomics research
- Identifying model organisms such as Mus musculus or Saccharomyces cerevisiae across publications to map experimental findings to specific species.
- Biomedical knowledge graph construction
- Parsing entities from unstructured biomedical texts to link species references with related biological entities and experimental contexts.
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 SpeciesDetect
Strengths and limits
Good at
- Trained on the curated LINNAEUS dataset focused directly on taxonomic and species entity recognition.
- Detects both scientific binomial nomenclature and common names of organisms across diverse texts.
- Achieves high recall (0.84) and overall token accuracy (0.98) on species-identification benchmarks.
Watch out for
- Has lower precision (0.72) and overall F1 score (0.78) compared to larger OpenMed species-detection models on the LINNAEUS benchmark.
- Entity outputs are restricted strictly to species mentions (B-SPECIES and I-SPECIES) and do not classify other biomedical entities.
- Does not directly perform normalisation to NCBI Taxonomy IDs as part of its output pipeline, producing only token spans and labels.
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 SpeciesDetect 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 SpeciesDetect 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 types can this model identify?
The model identifies species and organism mentions using two BIO tags: B-SPECIES for the start of a species name and I-SPECIES for inside or continuation tokens.
What corpus was used to train the model?
It was trained on the LINNAEUS corpus, a specialised biomedical benchmark for organism and species name recognition normalised to NCBI Taxonomy identifiers.
How does this model compare to other OpenMed species detection models?
With an F1 score of 0.78, it trails larger variants such as OpenMed-NER-SpeciesDetect-PubMed-335M (F1 of 0.9649) and PubMed-109M (F1 of 0.9543) on the LINNAEUS benchmark.
What input format does the model take?
The model accepts standard English biomedical or clinical text and outputs token labels alongside confidence scores.
How to cite it
Maziyar Panahi. OpenMed NER: Open-Source, Domain-Adapted State-of-the-Art Transformers for Biomedical NER Across 12 Public Datasets. arXiv:2508.01630 [cs.CL], 2025.
Written from OpenMed SpeciesDetect'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.