Skip to content

Latest commit

 

History

History
29 lines (19 loc) · 824 Bytes

README.md

File metadata and controls

29 lines (19 loc) · 824 Bytes

PokémonGAN

A generative adversarial network that creates new Pokémon from a random noise vector.

Generated samples:

Training progress:

Creating dataset (optional)

  • Raw images are located under the raw_dataset directory
  • To create training_dataset, run python image_preprocessing.py

Training

  • Open PokemonGAN.ipynb in Colab and follow instructions in the notebook.
  • Root folder: My Drive/PokemonGAN
  • Training samples generated by vis_noise.npy will save to PokemonGAN/training_samples.
  • Weights will save to PokemonGAN/weights

Visualizing training

  • Once training is complete, download the training_samples directory from your Drive
  • Run python make_training_GIF.py