/deepdreamer

Framework for deepdreaming images with a customizable call stack.

Primary LanguagePythonOtherNOASSERTION

purpose

deepdreamer framework for doing parameterized experiments based on googles deepdream ipython notebook.

requirements

pycaffe. caffe has its own pile of requirements, so if you can get that to work, you should be able to use this just fine.

example

First, make a folder called 'models', then copy the bvlc_googlenet model data (from caffe) to the models folder, under a directory called "bvlc_googlenet". Alternatively, you can edit the sample_config .json files to point to where you currently have the model stored.

python run_dream.py -c sample_configs/simple_demo.json
python run_dream.py -c sample_configs/guide_demo.json

You'll now have deep dreamed images in the folder 'output'.

todo

Lots of things. Like document the full call interface for proc.Proc.process_image. It's a work in progress. Don't be too hard on it right now.

license

See LICENSE file