Conditional-Glow

This code is a Python implementation of the conditional-Glow introduced in the paper

"Structured Output Learning with Conditional Generative Flows". You Lu and Bert Huang. AAAI 2020.

Note: This code is used for the experiments of binary segmentation on the Weizmann Horse dataset. Some parts of the code are adapted from chaiyujin, and openai.

Requirements:

This code was tested using the the following libraries.

• Python 3.6.7
• Numpy 1.14.6
• Pytorch 1.2.0
• Pillow 5.3.0
• skimage 0.16.2

Running

• Download the dataset from here.
• Rename the forlders /rgb and /figure_ground to be /images, and /labels, respectively.
• Within the same folder, create files train.txt, valid.txt, and test.txt, which contain the names of images for training, validation, and test, respectively.
• Configure the parameters in the shell script train_cglow.sh
• In the terminal, run ./train_cglow.sh

Contact

Feel free to send me an email, if you have any questions or comments.