pylearn: dataset.py comparison

Added MinibatchIterator

comparison

equal deleted inserted replaced

-:f7dcfb5f9d5b
+:8039918516fe
 def __getslice__(self,*slice_args):
 """dataset[i:j] returns the subdataset with examples i,i+1,...,j-1."""
 raise NotImplementedError
+def minibatches(self,minibatch_size):
+"""Return an iterator for the dataset that goes through minibatches of the given size."""
+return MinibatchIterator(self,minibatch_size)
+class MinibatchIterator(object):
+"""
+Iterator class for FiniteDataSet that can iterate by minibatches
+(sub-dataset of consecutive examples).
+"""
+def __init__(self,dataset,minibatch_size):
+assert minibatch_size>0 and minibatch_size<len(dataset)
+self.dataset=dataset
+self.minibatch_size=minibatch_size
+self.current=-minibatch_size
+def __iter__(self):
+return self
+def next(self):
+self.current+=self.minibatch_size
+if self.current>=len(self.dataset):
+self.current=-self.minibatchsize
+raise StopIteration
+return self.dataset[self.current:self.current+self.minibatchsize]
 # we may want ArrayDataSet defined in another python file
 import numpy
 class ArrayDataSet(FiniteDataSet):
 # copy the field here
 result[:,slice(c,slice_width)]=self.data[field_slice]
 c+=slice_width
 return result
 return self.data

Mercurial > pylearn