Keras: Bug: ImageDataGenerator for multiple output and single input fails

Created on 13 Sep 2016 · 4Comments · Source: keras-team/keras

The image data generator image.py fails when I tried to train a single input and multiple output model with data augmentation by saying 'the shape of X and y must be the same.

But, for multiple output model, I supply a list of ndarrays and since the input is a single input, I supplied just X as is ndarray. I locally fixed it by checking to see if y is a list and bypass the exception and it works just fine.

stale

Source

esube

Most helpful comment

This maybe closed due to stale state, but I believe the bug is still there.

nylocx on 4 Oct 2017

👍6

All 4 comments

I have met the same problem before. And I solved it by modifying some source code.
In , NumpyArrayIterator class can not handle multiple output.
Below code is the modification to solve the problem of ImageDataGenerator for multiple output.

Problematic code in NumpyArrayIterator class
batch_y = self.y[index_array]
This code is not appropriate for multiple output indexing.
I modified the class like the below. This is a temporal solution. I hope this solution works fine with you. And I think the bug: ImageDataGenerator for multiple output and multiple output should be fixed in keras source officially.
`
class NumpyArrayIterator(Iterator):

def __init__(self, X, y, image_data_generator,
             batch_size=32, shuffle=False, seed=None,
             dim_ordering='default',
             save_to_dir=None, save_prefix='', save_format='jpeg'):
    if y is not None and len(X) != len(y):
        raise Exception('X (images tensor) and y (labels) '
                        'should have the same length. '
                        'Found: X.shape = %s, y.shape = %s' % (np.asarray(X).shape, np.asarray(y).shape))
    if dim_ordering == 'default':
        dim_ordering = K.image_dim_ordering()
    self.X = X
    self.y = y
    self.image_data_generator = image_data_generator
    self.dim_ordering = dim_ordering
    self.save_to_dir = save_to_dir
    self.save_prefix = save_prefix
    self.save_format = save_format
    super(NumpyArrayIterator, self).__init__(X.shape[0], batch_size, shuffle, seed)

def next(self):
    # for python 2.x.
    # Keeps under lock only the mechanism which advances
    # the indexing of each batch
    # see http://anandology.com/blog/using-iterators-and-generators/
    with self.lock:
        index_array, current_index, current_batch_size = next(self.index_generator)
    # The transformation of images is not under thread lock so it can be done in parallel
    batch_x = np.zeros(tuple([current_batch_size] + list(self.X.shape)[1:]))
    for i, j in enumerate(index_array):
        x = self.X[j]
        x = self.image_data_generator.random_transform(x.astype('float32'))
        x = self.image_data_generator.standardize(x)
        batch_x[i] = x
    if self.save_to_dir:
        for i in range(current_batch_size):
            img = array_to_img(batch_x[i], self.dim_ordering, scale=True)
            fname = '{prefix}_{index}_{hash}.{format}'.format(prefix=self.save_prefix,
                                                              index=current_index + i,
                                                              hash=np.random.randint(1e4),
                                                              format=self.save_format)
            img.save(os.path.join(self.save_to_dir, fname))
    if self.y is None:
        return batch_x
    '''
    batch_y = self.y[index_array]
    '''
    batch_y = {}
    for i in range(len(self.y)):
        name = list(self.y.keys())[i]
        batch_y[name] = self.y[name][index_array]

    return batch_x, batch_y