Runs under Python 3.5 and TensorFlow 1.6+
- If you don't already have it, git clone TensorFlow models.
- Copy the datasets and preprocessing folders over from models/reasearch/slim into the directory with segmentation_vgg_16.py etc.
- Create a directory called slim_pretained and download and unpack vgg_16.ckpt and vgg_19.ckpt into it. A table of checkpoints locations can be found in the slim README Pre-Trained Models section. For example vgg_16_2016_08_28.tar.gz, and vgg_19_2016_08_28.tar.gz are the download locations.
- Run segmentation_vgg_16.py
