This is a Simple Implementation of Existing DC GANS (http://carpedm20.github.io/faces/) but in Jupyter Notebook version for easy visualization of Errors and Tensor Dimensions.
To Run this File just install Requirements.txt and give the Path of the Training Data correctly in Jupyter Notebook and it will generate images which can be seen in Tensorboard
- git clone https://github.com/sauradip/DC-GANS-Tensorflow-Jupyter-Notebook.git
- cd DC-GANS-Tensorflow-Jupyter-Notebook/
- To do this simply exeute pip install -r requirements.txt
- Download Training data in a folder called Data and change path path = 'data/' in 2nd cell of ipython
- Run the Code in Cloud or Good GPU with atleast 1 million Training Data for good results and for visualisation run tensorboard --logdir=logs1/