The folder containing the 2D, 2.5D and 3D network implementations for the purpose of training
Used to combine files of levels already converted to the internal ConvNet data structure and generates the properties file needed for training network
Counts numer of instances of various levels
Calculates every state possible for the maps given to it. and then prints out 1% of them
Levels used for training, consists of four different level types explained in the implementation section of the report.
Restores a level from it's convnet structure to it's original form
calculates the manhattan distance for any convnet formattet level at any of the paths set by the program argument
tries to solve a level using the 3D network trained by classifier