| #### How to use | |
| You can use this model with Transformers *pipeline* for NER. | |
| ```python | |
| from transformers import AutoTokenizer, AutoModelForTokenClassification | |
| from transformers import pipeline | |
| tokenizer = AutoTokenizer.from_pretrained("Holako/NER_model_holako") | |
| model = AutoModelForTokenClassification.from_pretrained("Holako/NER_model_holako") | |
| nlp = pipeline("ner", model=model, tokenizer=tokenizer) | |
| example = "اسمي احمد" | |
| ner_results = nlp(example) | |
| print(ner_results) | |
| ``` | |
| #### Limitations and bias | |
| This model is limited by its training dataset of entity-annotated news articles from a specific span of time. This may not generalize well for all use cases in different domains. | |
| ======= | |
| #### Limitations and bias | |
| This model is limited by its training dataset of entity-annotated news articles from a specific span of time. This may not generalize well for all use cases in different domains. | |
| ## Training data | |
| Language|Dataset | |
| -|- | |
| Arabic | [ANERcorp](https://camel.abudhabi.nyu.edu/anercorp/) | |