Amazing Signts - Tensorflow 2.0 Implimentation of DCGAN

Training

• first time

    python dcgan_fix_512_512.py --train --iteration 1000 --batch_size 8 --data_dir /path/to/your/images
  

• continue last training (Assuming alreay run 1000 iterations)

    python dcgan_fix_512_512.py --train --iteration 1000 --batch_size 8 --data_dir /path/to/your/images --restore_model --pre_low_step 1000
  

Generate 8 Images

python dcgan_fix_512_512.py --batch_size 8 --result_dir /path/to/your/results

Results

around 8000 mountain images downloaded form Unsplash https://drive.google.com/open?id=1256PUvIb3K1e8yx5HueTry7UedJUFgmK

Demo Website

http://amazingsignts.ngrok.io

Weights

https://drive.google.com/open?id=1DppoNi9fswsGfwOQbDiYiiKjr1DnuVAX

Data

• around 8000 mountain images downloaded form Unsplash
• around 5000 building images downloaded form Unsplash
• around 4000 desert images downloaded form Unsplash
• around 3000 lake images downloaded form Unsplash
• around 1500 waterfall images downloaded form Unsplash