jburnford/dyslexic-writer-qwen3-0.6b
The jburnford/dyslexic-writer-qwen3-0.6b is a 0.8 billion parameter Qwen3-based causal language model developed by jburnford, fine-tuned for spelling and grammar correction. With a context length of 32768 tokens, this model is specifically optimized to assist dyslexic writers by correcting errors. It achieves an Exact Match Accuracy of 78.8% and an Error Fix Rate of 71.2%, making it suitable for mobile and embedded applications requiring efficient text correction.
Loading preview...
Overview
The jburnford/dyslexic-writer-qwen3-0.6b is a 0.8 billion parameter model based on the Qwen3 architecture, specifically fine-tuned by jburnford for spelling and grammar correction. It is designed to assist dyslexic writers by accurately identifying and correcting textual errors. The model was trained on approximately 495,000 examples, including word pairs, sentence corrections, and paragraph-level error injection from synthetic stories.
Key Capabilities
- Spelling and Grammar Correction: Optimized to fix common spelling and grammatical errors.
- High Preservation Rate: Achieves a 98.7% No-Error Preservation rate, ensuring text without errors remains largely unchanged.
- Performance Metrics: Demonstrates an Exact Match Accuracy of 78.8% and an Error Fix Rate of 71.2%.
- Efficient Size: At 0.8 billion parameters, it is suitable for resource-constrained environments.
Good For
- Dyslexic Writers: Provides targeted assistance for individuals with dyslexia.
- Mobile and Embedded Applications: Its compact size (0.6B parameters) makes it ideal for deployment on devices with limited computational resources.
- Automated Text Correction: Can be integrated into applications requiring efficient and accurate spelling and grammar checks.