readability-es-3class-paragraphs
| Property | Value |
|---|---|
| Parameter Count | 125M |
| License | CC-BY-4.0 |
| Language | Spanish |
| Task | Text Classification |
| F1 Score | 0.7881 |
What is readability-es-3class-paragraphs?
This is a specialized Spanish language model designed to assess text readability across three complexity levels: Basic, Intermediate, and Advanced. Built on the BERTIN RoBERTa architecture, it specifically analyzes paragraph-level text to determine its complexity level, making it valuable for educational content creators and language learning applications.
Implementation Details
The model is implemented using the RoBERTa architecture and fine-tuned on BERTIN's Spanish base model. It processes paragraph-level inputs and was trained on a diverse dataset including coh-metrix-esp corpus, website scrapes, and proprietary datasets like newsela-es and simplext.
- Architecture: RoBERTa-based BERTIN model
- Training Data: Mixed dataset approach including public and private sources
- Granularity: Paragraph-level analysis
- Performance: 0.7881 F1 macro average score
Core Capabilities
- Three-level classification of text complexity
- Paragraph-level readability assessment
- Spanish text processing optimization
- Integration with Common European Framework of Reference for Languages
Frequently Asked Questions
Q: What makes this model unique?
This model stands out for its specific focus on Spanish text readability assessment at the paragraph level, offering a three-class classification system aligned with educational standards. It's particularly valuable for content adaptation and educational material development.
Q: What are the recommended use cases?
The model is ideal for educational content creators, language learning platforms, and content adaptation services that need to assess and categorize Spanish text difficulty levels. It can be used for content simplification, curriculum development, and reading material selection.
