| # Dataset Card for Masked Floorplans (Alpaca-style) | |
| ## Summary | |
| This dataset contains structured JSON representations of residential floorplans where **one room has been removed (masked)**. Each example follows the **Alpaca-style instruction format** to fine-tune a language model to **predict the missing room's geometric attributes** based on a partial plan and a semantic conditioning signal. | |
| ## Format | |
| Each sample is a JSON object with the following fields: | |
| - **instruction**: A prompt like `"Predict the geometry of the missing room."` | |
| - **input**: A JSON-encoded partial floorplan with visible rooms and the type of the masked room. | |
| - **output**: The ground-truth geometry of the missing room, including polygon coordinates, area, orientation, and adjacency. | |
| ## Use Cases | |
| - Instruction-tuning of LLMs for geometric reasoning | |
| - Floorplan infill or layout prediction | |
| - Generative design based on semantic layout context |