To ensure the reproducibility of all our experiment results, we have provided the code used for our method on each task in this supplementary material.

1. First, please download the model weights and some data from this Google Drive link https://drive.google.com/drive/folders/1BBTirhcX4ewIuxHG6QBHmHbJ3vp_-qQ8?usp=sharing. Then place the model in the "./results/model" folder and the datasets in the "./datasets" folder.

2. To generate city layouts using our model, navigate to the "test" folder and execute the following command:

python block_gen.py

3. Here's a brief description of each folder in this repository:

classification: Contains the code for downstream classification tasks.
data_pre: Includes the data preprocessing code.
model: Houses the main model structure.
share: Contains some basic functions and classes shared across the codebase.
test: Contains test scripts.
	3dscene_gen.py: 2.5D generation
	block_gen.py: city layout generation
	complement_gen.py: city complementation
	infinite_gen.py: infinite generation
	roadcondition_gen.py: conditional generation based on road network
train: Contains train scripts.
