en_core_sci_lg_spacy38

This repository contains an unofficial compatibility repackaging of scispaCy en_core_sci_lg 0.5.4 for spaCy 3.8. It is not a newly trained model.

The pipeline supports English biomedical and scientific text processing, including tokenization, part-of-speech tagging, lemmatization, dependency parsing, entity mention detection, and static word vectors.

Changes

The trained weights, tokenizer, vocabulary, vectors, and pipeline component data are unchanged from the original model.

The following configuration values were corrected from strings to booleans so that the pipeline passes spaCy 3.8 configuration validation:

• normalize = "False" to normalize = false
• include_static_vectors = "True" to include_static_vectors = true

The distribution name and spaCy compatibility metadata were also updated to distinguish this build from the original distribution. This compatibility build was tested with spaCy 3.8.14.

Installation

pip install "en_core_sci_lg_spacy38 @ https://huggingface.co/omid-jf/en_core_sci_lg_spacy38/resolve/main/en_core_sci_lg_spacy38-0.5.4-py3-none-any.whl"

Usage

import spacy

nlp = spacy.load("en_core_sci_lg_spacy38")

text = (
    "Myeloid-derived suppressor cells (MDSCs) are immature myeloid cells "
    "with immunosuppressive activity. They accumulate in tumor-bearing "
    "mice and humans with different types of cancer, including "
    "hepatocellular carcinoma (HCC)."
)

doc = nlp(text)

for sentence in doc.sents:
    print(sentence.text)

for entity in doc.ents:
    print(entity.text, entity.label_)

Optional scispaCy components such as the abbreviation detector and entity linker are provided by the separate scispacy package and are not included in this pipeline.

Validation

This compatibility build was checked by loading the installed wheel under spaCy 3.8.14 and exercising every included pipeline component, batch processing, annotations, vectors, and document serialization. It was not retrained or independently re-evaluated on the original benchmark datasets.

Attribution and license

The original model was developed by the Allen Institute for Artificial Intelligence as part of scispaCy. Its listed sources are OntoNotes 5, Common Crawl, and GENIA 1.0.

This modified distribution remains licensed under the Creative Commons Attribution-ShareAlike 3.0 license.

If you use this model, please cite the original scispaCy publication:

@inproceedings{neumann-etal-2019-scispacy,
  title = {{S}cispa{C}y: Fast and Robust Models for Biomedical Natural Language Processing},
  author = {Neumann, Mark and King, Daniel and Beltagy, Iz and Ammar, Waleed},
  booktitle = {Proceedings of the 18th BioNLP Workshop and Shared Task},
  year = {2019},
  pages = {319--327},
  doi = {10.18653/v1/W19-5034}
}