Some sample and utility code for the Deep Learning Workshop in Stuttgart on 09.12.2017. Dependencies:
- Python 3.5/3.6
- Tensorflow (with Slim)
- Scipy
- Numpy
For Windows get Scipy and Numpy from (https://www.lfd.uci.edu/~gohlke/pythonlibs/) and install with
pip install <filename>
Some information on the different files and folders:
- loaders: some helpers for loading data are given.
- Dataset.py: contains an buffered multithreaded dataset loader for efficient loading of arbitrary datasets.
- Models: contains Googlenet code.
- trainer.py and trainer_onehot.py contain a simple model with all code that is need to train and debug a model.