nllg/TikZilla-8B
TikZilla-8B by nllg is an 8 billion parameter language model based on Qwen3-8B-Base, specifically fine-tuned for generating TikZ/LaTeX figures from natural language descriptions. It was trained using supervised fine-tuning on the DaTikZ-V4 dataset. This model excels at converting textual figure requirements into complete, valid LaTeX code, making it ideal for scientific document preparation and automated diagram generation.
Loading preview...
TikZilla-8B: Natural Language to TikZ/LaTeX Figure Generation
TikZilla-8B is an 8 billion parameter language model developed by nllg, specialized in generating TikZ/LaTeX figures directly from natural language descriptions. Built upon the Qwen3-8B-Base architecture, this model has undergone supervised fine-tuning (SFT) using the DaTikZ-V4 dataset, specifically targeting the creation of scientific and technical diagrams.
Key Capabilities
- TikZ/LaTeX Code Generation: Translates detailed natural language descriptions into complete and valid LaTeX documents containing TikZ figures.
- Scientific Figure Creation: Optimized for generating common scientific plots, diagrams, and visual representations.
- Contextual Understanding: Processes complex textual requirements to produce accurate and structured graphical code.
Good for
- Automating Figure Creation: Developers and researchers needing to quickly generate diagrams for papers, presentations, or reports without manual TikZ coding.
- Scientific Documentation: Enhancing workflows for academic writing by converting figure specifications into ready-to-use LaTeX code.
- Prototyping Visualizations: Rapidly prototyping visual ideas by describing them in plain text and getting immediate TikZ output.