SlipGURU Dipartimento di Informatica e Scienze dell'Informazione Università Degli Studi di Genova

Command-line Applications

This is the script already shown in the brief introductory tutorial, the file can be found in the examples directory of the source distribution.

Run the script with -h to get a list of possible options:

$ python paddle/examples/ -h
Usage: imagefile [options]

  -h, --help      show this help message and exit
  -n INT          limit the number of training patches to INT
  -p INT          set the size of the patches to INTxINT (default 12)
  -k INT          set the size of the dictionary to INT (default 200)
  --tau=FLOAT     set the weight of sparsity penalty to FLOAT (default 0.5)
  --npzfile=PATH  save the results to PATH in npz format
  --outpath=PATH  save to PATH three images with a sample of the training
                  patches, the dictionary atoms and their dual filters, all
                  arranged in tables
  --nrows=INT     set the number of rows of the tables to INT (default 10)
  --ncols=INT     set the number of columns of the tables to INT (default 20)

Table Of Contents