** Currently, you need the following version of theano to use the sparse implementation: https://bitbucket.org/ynd/theano/overview #----------------------------------------------------------------------------------------- SCRIPTS DESCRIPTIONS: #----------------------------------------------------------------------------------------- ** model.py contains all the tools to build, train and evaluate a model. ** datatools.py shows how to build the necessary data file (sparse matrices, dictionnaries), from the nlu/ data. ** simpleexpe.py is a simplified version of the training script (only using WordNet3.0 data), you can change directly the hard-coded parameters and run the code in the WakaBST folder using one of the two following commands: THEANO_FLAGS=floatX=float32,device=gpu python simpleexpe.py (for gpu use) or THEANO_FLAGS=floatX=float32,device=cpu python simpleexpe.py (for cpu use) ** utils.py is a simple script to load a model and create ranking lists. THEANO_FLAGS=floatX=float32,device=[cpu/gpu] python utils.py ** evaluation.py shows how to do the different evaluations procedure (mean rank, WSD, ...etc) THEANO_FLAGS=floatX=float32,device=[cpu/gpu] python evaluation.py resultfolder idjob evalnumber ** expeWAKA.py (sorry for the headache...) is the training script over all the different training sets using the jobman interface. #------------------------------------------------------------------------------------------ Contact: myusername ..at.. iro.umontreal.ca