| Property | Value |
|---|---|
| Framework | PyTorch/Flair |
| Task | Named Entity Recognition |
| Performance | 89.27% F1-Score |
| Dataset | Ontonotes |
What is ner-english-ontonotes?
The ner-english-ontonotes is a state-of-the-art Named Entity Recognition model developed using the Flair framework. It's designed to identify and classify 18 different types of entities in English text, including person names, organizations, dates, and monetary values. The model leverages sophisticated Flair embeddings combined with LSTM-CRF architecture to achieve high accuracy in entity detection.
Implementation Details
This model implements a sophisticated architecture combining Flair embeddings with an LSTM-CRF layer. It uses a stacked embedding approach that includes GloVe embeddings and both forward and backward contextual string embeddings, trained with a hidden size of 256 units.
- Utilizes stacked embeddings (GloVe + Flair forward/backward)
- LSTM-CRF architecture for sequence labeling
- Trained on the Ontonotes dataset
- Supports 18 distinct entity classes
Core Capabilities
- Recognition of person names, organizations, and locations
- Detection of temporal expressions (dates and times)
- Identification of monetary values and quantities
- Classification of work of art titles and product names
- Recognition of geopolitical entities and affiliations
Frequently Asked Questions
Q: What makes this model unique?
This model stands out for its comprehensive coverage of 18 entity types and high F1-score of 89.27% on the Ontonotes dataset. It combines the power of contextual string embeddings with traditional word embeddings, making it particularly effective for complex entity recognition tasks.
Q: What are the recommended use cases?
The model is ideal for applications requiring detailed entity extraction from English text, such as information extraction systems, document analysis tools, and automated content categorization. It's particularly useful when fine-grained entity classification is needed beyond basic person/organization/location recognition.