jamesjk4242/insighta-mandala-v13
The Insighta Mandala v13 by jamesjk4242 is a 4 billion parameter Qwen3-based model, fine-tuned specifically for generating structured JSON mandala learning plans. It supports both Korean and English inputs and excels at creating detailed, organized learning charts. This model is optimized for educational planning and structured output generation.
Loading preview...
Insighta Mandala v13: Structured Learning Plan Generation
Insighta Mandala v13 is a specialized 4 billion parameter language model, built upon the Qwen3-4B architecture and fine-tuned by jamesjk4242. Its core capability lies in generating structured JSON-formatted mandala learning plans, making it unique among general-purpose LLMs. The model supports both Korean and English, allowing for versatile application in different linguistic contexts.
Key Capabilities
- Mandala Learning Plan Generation: Specifically designed to create detailed, organized learning plans in a mandala chart format.
- Structured JSON Output: Outputs plans as clean, parseable JSON, facilitating integration into applications or workflows.
- Multilingual Support: Processes prompts and generates content in both Korean and English.
- Optimized for Specific Task: Fine-tuned using LoRA for high performance on its niche task, rather than broad general knowledge.
Good For
- Educational Tools: Developing applications that require automated generation of structured study plans.
- Personalized Learning: Creating customized learning roadmaps for students based on specific goals (e.g., "TOEFL 100점 만다라트 차트").
- Structured Data Generation: Any use case requiring a model to reliably output complex, nested JSON structures based on natural language prompts.
Quantized versions, including Q4_K_M.gguf (recommended for CPU), are available for efficient deployment.