- This project is the code of the report "LittleGAN: Conditional Facial Image Generation and Adjustment"
- This project use keras to build the model, use tensorflow as backend to train, test, evaluate, predict.
- We tested and run this project on ubuntu 16.04, cuda 9.0, cudnn 7.3.1, tensorflow 1.12.
- This project can be run on cpu or nvidia gpu, but we recommend using gpu to train it.
- We use the dataset "CelebA" to train model. please download it first from http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html
- Please crop the images to 128*128. We recommend to use this project HD-CelebA-Cropper.
- If you wanna more data or information about this project, please contact us.
- There is a demo: LittleGAN Live and Source Code