jburnford/dyslexic-writer-qwen3-1.7b
The jburnford/dyslexic-writer-qwen3-1.7b is a 2 billion parameter Qwen3-based causal language model fine-tuned by jburnford. It specializes in spelling and grammar correction, specifically optimized to assist dyslexic writers. This model achieves an 82.2% exact match accuracy and a 75.9% error fix rate, making it suitable for applications requiring precise text correction.
Loading preview...
Dyslexic Writer - Qwen3-1.7B Overview
This model, jburnford/dyslexic-writer-qwen3-1.7b, is a fine-tuned version of the Qwen/Qwen3-1.7B architecture, specifically developed by jburnford. Its primary purpose is to provide robust spelling and grammar correction, with a particular focus on aiding dyslexic writers. The model has been trained on approximately 495,000 examples, incorporating diverse data such as word pairs, sentence corrections, and paragraph-level error injection from synthetic stories.
Key Capabilities
- High Accuracy Correction: Achieves an 82.2% exact match accuracy and a 75.9% error fix rate for spelling and grammar.
- Error Preservation: Demonstrates a 98.9% no-error preservation rate, ensuring correct text remains unchanged.
- Optimized for Dyslexia: Specifically fine-tuned to address common writing challenges faced by dyslexic individuals.
- Efficient Size: At 1.7 billion parameters, it offers a balance between performance and resource usage.
Good For
- Automated Proofreading: Ideal for integrating into writing tools or applications that require automatic spelling and grammar checks.
- Assistive Technologies: Particularly beneficial for creating tools designed to support dyslexic users in their writing process.
- Text Preprocessing: Can be used to clean and standardize text data by correcting common errors before further processing.