# dmis-lab/biobert

Bioinformatics'2020: BioBERT: a pre-trained biomedical language representation model for biomedical text mining

Repository: https://github.com/dmis-lab/biobert
Canonical: https://ross.abutalabs.com/products/biobert
Homepage: http://doi.org/10.1093/bioinformatics/btz682
Language: Python
License: NOASSERTION
License Family: other
Last push: 2023-08-13T21:11:54+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2778, "days_push": 1116, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2204, forks 478 (observed 2026-08-28T04:06:25.896703+00:00)

## What it is
BioBERT is a pre-trained biomedical language representation model (based on BERT) with code for fine-tuning it on biomedical text mining tasks such as named entity recognition, relation extraction, and question answering. The repository distributes multiple versions of pre-trained weights trained on PubMed and PMC corpora, primarily for TensorFlow 1 with a separate PyTorch port.

## Use cases
- fine-tune a pretrained BERT model on biomedical text
- named entity recognition for diseases and drugs in PubMed abstracts
- extract relations between biomedical entities from scientific literature
- biomedical question answering over research text
- download BioBERT pretrained weights for transfer learning
- adapt a language model to domain-specific biomedical corpora

## When to choose
- you need a domain-adapted BERT checkpoint for biomedical NLP tasks
- you are reproducing the BioBERT paper's NER, RE, or QA benchmarks
- you want to fine-tune on PubMed/PMC-derived text with TensorFlow 1 or the PyTorch port
- you need a proven baseline model for biomedical text mining research

## When to avoid
- you need a modern, actively maintained stack (the code targets TensorFlow 1 and Python <= 3.7)
- your text is general-domain rather than biomedical (plain BERT or newer models fit better)
- you want a no-code entity recognizer rather than a fine-tuning codebase (use the linked BERN tool instead)
- you require a permissive license - the license is non-standard and must be reviewed

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, nlp, deep-learning, transformers
- domain: bioinformatics, healthcare, artificial-intelligence
- platform: python
- tags: bert, biomedical-text-mining, pretrained-model, named-entity-recognition, relation-extraction, question-answering, pubmed, tensorflow, fine-tuning, research, natural-language-processing, linux, gpu

## Member repositories
- dmis-lab/biobert (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:25.896703+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:46:43.510455+00:00, confidence not recorded.
  - readme: https://github.com/dmis-lab/biobert (fetched 2026-08-28T04:06:25.896703+00:00, sha 47adbcbb2f65)
- Data as of 2026-08-30T08:39:29.467469+00:00.