Materials for Thematic Map Generation System (code, data, trained weights)

This release contains the complete thematic map generation package, including source code, runtime scripts, training datasets, project-trained checkpoints, and demo inputs. Code and web/API workflow are provided under app/, with a headless reproducibility script in scripts/run_demo.py. Training datasets are included under datasets/Controlnet training data/ and datasets/LoRA training data/. Project-trained checkpoints are included under models/Controlnet/Colorization.ckpt and models/Lora/Map symbol.safetensors. Demo reproducibility inputs are included under data/presentation/ (province map GeoJSON + matched POI GeoJSON). To reproduce a deterministic baseline output, run: python scripts/run_demo.py --demo beijing --seed 42 --with-legend. This command (fixed seed = 42) generates a stable reference output for reviewer-side verification. The Stable Diffusion 1.5 base model is not distributed in this archive; please download it separately as documented in README.md.