dcgan_with_encoder

Implementation of DC-GAN with Encoder using TensorLayer2

Requirement

• Python3
• TensorFlow==2.0.0a0
• TensorLayer==2.1.0
• CPU or NVIDIA GPU + CUDA CuDNN

Train the Model

Clone this repo

git clone https://github.com/YHQpkueecs/dcgan_with_encoder.git

Create a folder data and download celebA dataset.

Run train-gd.py train-e.py to train the generator+discriminator and encoder sequentially.

python3 train-gd.py
python3 train-e.py

Modify and run test.py for evaluation.

Results

Randomly generate 64 faces

Interpolate between two randomly generated faces

Encoding and interpolation for two selected images