| | --- |
| | tags: |
| | - spacy |
| | - token-classification |
| | language: |
| | - en |
| | model-index: |
| | - name: en_grant |
| | results: |
| | - task: |
| | name: NER |
| | type: token-classification |
| | metrics: |
| | - name: NER Precision |
| | type: precision |
| | value: 0.8150708215 |
| | - name: NER Recall |
| | type: recall |
| | value: 0.7125309559 |
| | - name: NER F Score |
| | type: f_score |
| | value: 0.760359408 |
| | --- |
| | ## Introduction |
| |
|
| | Three variants of the model is built with Spacy3 for grant applications. |
| | A simple named entity recognition custom model from scratch with annotation tool prodi.gy. |
| | Github info: https://github.com/RaThorat/ner_model_prodigy |
| | The most general model is 'en_grantss'. The model en_ncv is more suitable to extract entities from narrative CV's. |
| | The model en_grant is the first model in the series. |
| | |
| | | Feature | Description | |
| | | --- | --- | |
| | | **Name** | `en_grant` | |
| | | **Version** | `0.0.0` | |
| | | **spaCy** | `>=3.4.3,<3.5.0` | |
| | | **Default Pipeline** | `tok2vec`, `ner` | |
| | | **Components** | `tok2vec`, `ner` | |
| | | **Vectors** | 0 keys, 0 unique vectors (0 dimensions) | |
| | | **Sources** | research grant applications | |
| | | **License** | n/a | |
| | | **Author** | [Rahul Thorat]() | |
| |
|
| | ### Label Scheme |
| |
|
| | <details> |
| |
|
| | <summary>View label scheme (18 labels for 1 components)</summary> |
| |
|
| | | Component | Labels | |
| | | --- | --- | |
| | | **`ner`** | `ACTIVITY`, `DISCIPLINE`, `GPE`, `JOURNAL`, `KEYWORD`, `LICENSE`, `MEDIUM`, `METASTD`, `MONEY`, `ORG`, `PERSON`, `POSITION`, `PRODUCT`, `RECOGNITION`, `REF`, `REPOSITORY`, `WEBSITE`, `YEAR` | |
| |
|
| | </details> |
| |
|
| | ### Accuracy |
| |
|
| | | Type | Score | |
| | | --- | --- | |
| | | `ENTS_F` | 76.04 | |
| | | `ENTS_P` | 81.51 | |
| | | `ENTS_R` | 71.25 | |
| | | `TOK2VEC_LOSS` | 9725604.63 | |
| | | `NER_LOSS` | 930155.13 | |
