academic-figure-generation

The pipeline converts method text and captions into publication-quality figures through a multi-agent workflow (Retriever, Planner, Stylist, Visualizer, Critic) that produces multiple candidate visuals for selection.

Yes, the pipeline transforms method text and captions into architecture diagrams, framework visuals, and methodology illustrations by processing the content through configurable model settings.

Yes, you configure model parameters via configs/model_config.yaml to adjust the multi-agent pipeline settings before generating publication-quality academic figures.

The best way to produce multiple figure candidates is running the generate.py script, which applies a multi-agent pipeline to method text to output multiple figure candidates for comparison.

Yes, Pillow is a required dependency for the pipeline to process and render publication-quality academic figures from methodological text and target captions.