Tensorflow implementation of Bridging the Gap between Label- and Reference-based Synthesis in Multi-attribute Image-to-Image Translation.

Overview architecture

• CelebA

• Prerequisites
  • Tensorflow 1.15
  • Python 2.x with matplotlib, numpy and scipy
• Dataset
  • CelebA
  • Images should be placed in ./data/CelebA/img_align_celeba
  • tfrecords file should be placed in ./data/CelebA/celeba_tfrecords

Exemplar commands are listed here for a quick start.

• prepare dataset to product tfrecords file

  python data_noise_in.py 
  

• To train with size of 128 X 128

  python train_arch9.py --experiment_name "file_name" --gpu "gpu_num"
  

• Example of test

  python  test_arch9.py --experiment_name "file_name" --gpu "gpu_num"

If this work is useful for your research, please consider citing: