ncbi-nlp/BioSentVec
BioWordVec & BioSentVec: pre-trained embeddings for biomedical words and sentences
Trained on 4.9 billion tokens from PubMed literature and MIMIC-III clinical notes, this project provides fastText word vectors (200-dim) and sent2vec sentence embeddings (700-dim) optimized for biomedical NLP tasks. The models handle out-of-vocabulary terms through fastText's subword approach and are evaluated on domain-specific similarity benchmarks (MayoSRS, BIOSSES, MedSTS), outperforming general-purpose embeddings like Universal Sentence Encoder on clinical text.
611 stars. No commits in the last 6 months.
Stars
611
Forks
99
Language
Jupyter Notebook
License
—
Category
Last pushed
Aug 15, 2023
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/nlp/ncbi-nlp/BioSentVec"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
avidale/compress-fasttext
Tools for shrinking fastText models (in gensim format)
dselivanov/text2vec
Fast vectorization, topic modeling, distances and GloVe word embeddings in R.
vzhong/embeddings
Fast, DB Backed pretrained word embeddings for natural language processing.
dccuchile/spanish-word-embeddings
Spanish word embeddings computed with different methods and from different corpora
ibrahimsharaf/doc2vec
:notebook: Long(er) text representation and classification using Doc2Vec embeddings