ner-vietnamese-electra-base
| Property | Value |
|---|---|
| Parameter Count | 133M |
| Model Type | Token Classification |
| Architecture | ELECTRA-based |
| Downloads | 134,125 |
| Overall F1 Score | 92.14% |
What is ner-vietnamese-electra-base?
This is a specialized Named Entity Recognition (NER) model fine-tuned on the VLSP 2018 dataset, built upon the ELECTRA architecture specifically for Vietnamese text analysis. The model demonstrates exceptional performance in identifying and classifying named entities, with particularly strong results in person name detection (96.64% F1) and location recognition (93.65% F1).
Implementation Details
The model utilizes the Transformers pipeline for seamless integration, trained with Adam optimizer using carefully tuned hyperparameters (learning rate: 5e-05, batch size: 16). It processes Vietnamese text to identify four main entity types: Location, Person, Organization, and Miscellaneous.
- Location Entity Detection: 93.65% F1 score
- Person Name Recognition: 96.64% F1 score
- Organization Detection: 88.33% F1 score
- Overall Accuracy: 99.07%
Core Capabilities
- High-precision Vietnamese named entity recognition
- Efficient processing with 133M parameters
- Support for multiple entity types
- Easy integration via Transformers pipeline
- Optimized for practical applications
Frequently Asked Questions
Q: What makes this model unique?
The model's exceptional performance on Vietnamese text, particularly its high F1 scores across different entity types, makes it a standout choice for Vietnamese NER tasks. Its balanced performance across entity types demonstrates robust training and optimization.
Q: What are the recommended use cases?
This model is ideal for applications requiring Vietnamese text analysis, including information extraction, document processing, and automated content analysis. It's particularly effective for tasks requiring accurate identification of people, locations, and organizations in Vietnamese text.
