A personal project to learn Autoencoders
I am interested in the use of Neural Networks to generate new things from an exsisting style I choose the Pokemon images because they were all the same sizes, and yet still had a distinct style and I could find a good number of them
requirements:
pip3 install tensorflow
pip3 install bokeh
git lfs pull (the grey_scale_larger model is too large for github)
run with:
bokeh serve --show ./bokeh_server.py
Then adjust the weights and click the Generate Pokemon button to see what changes you have made