Required packages : -> Tensorflow 0.10 -> Python 3.5 -> Scikit-learn 0.17 -> Numpy -> Scipy The code was developped under this versions of the packages, it might be incompatible with some newer versions due to the depreciation of some functions.
You can download the datasets by using: the import.sh script from the bash subfloder In order to specify the Dataset, you should add an option next to the script 0 : 100K 1 : 1M 2 : 10M (For example ./import.sh 1), Then you can choose the configuration by changing the first line on the main.py file. For example: from config.autoencoder.U_1M import Experiment This runs the Autoencoder for 1M dataset under U setting (user setting) from config.stability.V_1M import Experiment This runs the Stability for 1M dataset under V setting (movie setting)
You can also change the experiment parameters under the config/ subfolder.