| --- |
| language: en |
| library_name: spacy |
| pipeline_tag: token-classification |
| tags: |
| - spacy |
| - token-classification |
| - ner |
| extra_gated_prompt: "Access is granted to collaborators of the NER-FOR-PROPOSAL-GRAPHRAG project." |
| --- |
| |
| # en_proposal_ner_trf |
| |
| spaCy NER pipeline (`roberta-base` transformer + NER) for user queries about |
| research proposals, used by |
| [NER-FOR-PROPOSAL-GRAPHRAG](https://github.com/xhris12/NER-FOR-PROPOSAL-GRAPHRAG). |
| |
| Labels: `CAREER_STAGE`, `COUNTRY`, `DEPARTMENT`, `FRAMEWORK`, `INFRA_KEYWORD`, |
| `KEYWORD`, `NATIONALITY`, `ORG`, `ORG_TYPE`, `PERSON`, `PROJECT_NAME_FRAGMENT`, |
| `PROPOSAL`, `SEARCH_PHRASE`, `TOPIC`. |
|
|
| ## Usage |
|
|
| Requires `spacy>=3.8.16,<3.9` and `spacy-transformers>=1.4,<1.5`, plus access |
| to this repo and a Hugging Face token (`hf auth login` or `HF_TOKEN`). |
|
|
| ```python |
| import spacy |
| from huggingface_hub import snapshot_download |
| |
| nlp = spacy.load(snapshot_download("CKris21/en_proposal_ner_trf")) |
| doc = nlp("Proposals by Maria Rossi under Horizon Europe") |
| print([(ent.text, ent.label_) for ent in doc.ents]) |
| ``` |
|
|
