Paper link

This repository contains code for the paper: Content Aware Layout Generation link

Setup

conda create -p ./env
source activate ./env
pip install tensorflow-gpu pillow scipy tf_slim

Store the dataset, example, sample, validation directories inside a data directory outside the LayoutNet directory.

Training

• python train.py
• Results are generated in the folder ./example/*.png
• Models are saved in the folder ./log/***

Testing

• Generate one random noise vector (60*128).
• Change the testing folder in inputs() from ./sample/ to ./validation/
• Select the model name in testing()
• python test.py