allenai/scibert
A BERT language model pretrained on scientific publications.
Not currently ranked — collecting fresh signals.
star history
SciBERT is a pretrained contextualized embeddings model based on BERT architecture, trained on a large corpus of scientific papers from Semantic Scholar. It is designed to improve performance on downstream NLP tasks in the scientific domain such as named entity recognition, relation extraction, and citation intent classification.
Frequently asked
- What is allenai/scibert?
- A BERT language model pretrained on scientific publications.
- Is scibert open source?
- Yes — allenai/scibert is open source, released under the Apache-2.0 license.
- What language is scibert written in?
- allenai/scibert is primarily written in Python.
- How popular is scibert?
- allenai/scibert has 1.7k stars on GitHub.
- Where can I find scibert?
- allenai/scibert is on GitHub at https://github.com/allenai/scibert.