/simplified_pix2pixHD

Simplified repo of pix2pixHD for quick prototyping(unofficial)

Primary LanguageJupyter NotebookMIT LicenseMIT

Simplified pix2pixHD repo(Unofficial)

The simplified version of High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs code for quick prototyping and experiments.

Example

How to use this repo?

  1. Download sources from github: git clone https://github.com/Hramchenko/simplified_pix2pixHD.git
  2. Open simplified_pix2pixHD.ipynb in Jupyter.
  3. Download images to data folder(for example facades.tar.gz from here).
  4. Modify the Dataset class, the generator and the discriminator as you need.
  5. Run Jupyter cells for training.

Requirements

  1. pytorch
  2. opencv-python
  3. numpy
  4. matplotlib
  5. tqdm