New

• We clean up the code to make it more readable! The old version is here: v1.

---

     

---

AttGAN
_{TIP Nov. 2019, arXiv Nov. 2017}

TensorFlow implementation of AttGAN: Facial Attribute Editing by Only Changing What You Want

Related

• Other implementations of AttGAN

  • AttGAN-PyTorch by Yu-Jing Lin

  • AttGAN-PaddlePaddle by ceci3 and zhumanyu (AttGAN is one of the official reproduced models of PaddlePaddle)

• Closely related works

  • An excellent work built upon our code - STGAN (CVPR 2019) by Ming Liu

  • Changing-the-Memorability (CVPR 2019 MBCCV Workshop) by acecreamu

  • Fashion-AttGAN (CVPR 2019 FSS-USAD Workshop) by Qing Ping

• An unofficial demo video of AttGAN by 王一凡

Exemplar Results

• See results.md for more results, we try higher resolution and more attributes (all 40 attributes!!!)

• Inverting 13 attributes respectively

  from left to right: Input, Reconstruction, Bald, Bangs, Black_Hair, Blond_Hair, Brown_Hair, Bushy_Eyebrows, Eyeglasses, Male, Mouth_Slightly_Open, Mustache, No_Beard, Pale_Skin, Young

  

Usage

• Environment

  • Python 3.6

  • TensorFlow 1.15

  • OpenCV, scikit-image, tqdm, oyaml

  • we recommend Anaconda or Miniconda, then you can create the AttGAN environment with commands below

    conda create -n AttGAN python=3.6
    
    source activate AttGAN
    
    conda install -c anaconda tensorflow-gpu=1.15
    
    conda install -c anaconda opencv
    
    conda install -c anaconda scikit-image
    
    conda install -c anaconda tqdm
    
    conda install -c conda-forge oyaml

• Data Preparation

  • CelebA-unaligned (10.2GB, higher quality than the aligned data)

    • download the dataset

      • img_celeba.7z (move to ./data/img_celeba/img_celeba.7z): Google Drive or Baidu Netdisk

      • annotations.zip (move to ./data/img_celeba/annotations.zip): Google Drive

    • unzip and process the data

      7z x ./data/img_celeba/img_celeba.7z/img_celeba.7z.001 -o./data/img_celeba/
      
      unzip ./data/img_celeba/annotations.zip -d ./data/img_celeba/
      
      python ./scripts/align.py

• Run AttGAN

  • NOTICE: if you create a new conda environment, remember to activate it before any command

    source activate AttGAN

  • training (see examples.md for more training commands)

    CUDA_VISIBLE_DEVICES=0 \
    python train.py \
    --load_size 143 \
    --crop_size 128 \
    --model model_128 \
    --experiment_name AttGAN_128

  • testing

    • single attribute editing (inversion)

      CUDA_VISIBLE_DEVICES=0 \
      python test.py \
      --experiment_name AttGAN_128

    • multiple attribute editing (inversion) example

      CUDA_VISIBLE_DEVICES=0 \
      python test_multi.py \
      --test_att_names Bushy_Eyebrows Pale_Skin \
      --experiment_name AttGAN_128

    • attribute sliding example

      CUDA_VISIBLE_DEVICES=0 \
      python test_slide.py \
      --test_att_name Pale_Skin \
      --test_int_min -2 \
      --test_int_max 2 \
      --test_int_step 0.5 \
      --experiment_name AttGAN_128

  • loss visualization

    CUDA_VISIBLE_DEVICES='' \
    tensorboard \
    --logdir ./output/AttGAN_128/summaries \
    --port 6006

  • convert trained model to .pb file

    python to_pb.py --experiment_name AttGAN_128

• Using Trained Weights

  • alternative trained weights (move to ./output/*.zip)

    • AttGAN_128.zip (987.5MB)

      • including G, D, and the state of the optimizer

    • AttGAN_128_generator_only.zip (161.5MB)

      • G only

    • AttGAN_384_generator_only.zip (91.1MB)

  • unzip the file (AttGAN_128.zip for example)

    unzip ./output/AttGAN_128.zip -d ./output/

  • testing (see above)

• Example for Custom Dataset

  • AttGAN-Cartoon

Citation

If you find AttGAN useful in your research work, please consider citing:

@ARTICLE{8718508,
author={Z. {He} and W. {Zuo} and M. {Kan} and S. {Shan} and X. {Chen}},
journal={IEEE Transactions on Image Processing},
title={AttGAN: Facial Attribute Editing by Only Changing What You Want},
year={2019},
volume={28},
number={11},
pages={5464-5478},
keywords={Face;Facial features;Task analysis;Decoding;Image reconstruction;Hair;Gallium nitride;Facial attribute editing;attribute style manipulation;adversarial learning},
doi={10.1109/TIP.2019.2916751},
ISSN={1057-7149},
month={Nov},}