Python code to optimize Deep neural network created with pylearn2 using Spearmint.
To setup Pylearn2 and Spearmint please follow the instructions given on the respective page.
Additionally the ''Data Analysis Framework: ROOT'' including python bindings needs to be installed.
To install just clone the repository to your home directory
git clone https://github.com/MarkusSpanring/hepDnn
The files in the folders
- pylearn2
- Spearmint
in the master branch need to be copied into the respective folder of the pylearn2 and Spearmint installation directories.
The files main.py and launcher.py in the spearmint folder need to be merged with the respective file in this folder. The modifications are necessery to run optimize_DNN.py and scheduler.py.
The file higgs_dataset.py in the pylearn2/datasets needs to be copied to the respective installation directory of Pylearn2.
The file mlp.py in the pylearn2/models folder needs to be merged with the respective file in the installation directory.