view datasets/dataset.py @ 261:6d16a2bf142b

important bug fix in pipeline corresponding to an iterator always giving 0 complexity, need to redo 10% of the P07 dataset
author Xavier Glorot <glorotxa@iro.umontreal.ca>
date Wed, 17 Mar 2010 16:41:16 -0400
parents d6672a7daea5
children
line wrap: on
line source

from dsetiter import DataIterator

class DataSet(object):
    def test(self, batchsize, bufsize=None): 
        r"""
        Returns an iterator over the test examples.

        Parameters
          batchsize (int) -- the size of the minibatches
          bufsize (int, optional) -- the size of the in-memory buffer,
                                     0 to disable.
        """
        return self._return_it(batchsize, bufsize, self._test)

    def train(self, batchsize, bufsize=None):
        r"""
        Returns an iterator over the training examples.

        Parameters
          batchsize (int) -- the size of the minibatches
          bufsize (int, optional) -- the size of the in-memory buffer,
                                     0 to disable.
        """
        return self._return_it(batchsize, bufsize, self._train)

    def valid(self, batchsize, bufsize=None):
        r"""
        Returns an iterator over the validation examples.

        Parameters
          batchsize (int) -- the size of the minibatches
          bufsize (int, optional) -- the size of the in-memory buffer,
                                     0 to disable.
        """
        return self._return_it(batchsize, bufsize, self._valid)

    def _return_it(batchsize, bufsize, data):
        r"""
        Must return an iterator over the specified dataset (`data`).

        Implement this in subclassses.
        """
        raise NotImplemented