Error while fetching MNIST data

[kulshekhar]

In the exercise on handwritten digit recognition (in the 'Intro to TFLearn' lesson), I'm getting the following error when trying to execute the code that loads the MNIST data:

Extracting mnist/train-images-idx3-ubyte.gz
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-3-bb487e4cbb6a> in <module>()
      1 # Retrieve the training and test data
----> 2 trainX, trainY, testX, testY = mnist.load_data(one_hot=True)

/home/user/anaconda3/envs/tensorflow/lib/python3.5/site-packages/tflearn/datasets/mnist.py in load_data(one_hot)
     13 
     14 def load_data(one_hot=False):
---> 15     mnist = read_data_sets("mnist/", one_hot=one_hot)
     16     return mnist.train.images, mnist.train.labels, mnist.test.images, mnist.test.labels
     17 

/home/user/anaconda3/envs/tensorflow/lib/python3.5/site-packages/tflearn/datasets/mnist.py in read_data_sets(train_dir, fake_data, one_hot)
    159     VALIDATION_SIZE = 5000
    160     local_file = maybe_download(TRAIN_IMAGES, train_dir)
--> 161     train_images = extract_images(local_file)
    162     local_file = maybe_download(TRAIN_LABELS, train_dir)
    163     train_labels = extract_labels(local_file, one_hot=one_hot)

/home/user/anaconda3/envs/tensorflow/lib/python3.5/site-packages/tflearn/datasets/mnist.py in extract_images(filename)
     48         rows = _read32(bytestream)
     49         cols = _read32(bytestream)
---> 50         buf = bytestream.read(rows * cols * num_images)
     51         data = numpy.frombuffer(buf, dtype=numpy.uint8)
     52         data = data.reshape(num_images, rows, cols, 1)

/home/user/anaconda3/envs/tensorflow/lib/python3.5/gzip.py in read(self, size)
    272             import errno
    273             raise OSError(errno.EBADF, "read() on write-only GzipFile object")
--> 274         return self._buffer.read(size)
    275 
    276     def read1(self, size=-1):

TypeError: only integer scalar arrays can be converted to a scalar index

I get this error when executing the code from the notebook and also while executing it in the REPL.

I spent some time on this but was unable to figure out what I could do to fix this. Any pointers would be very helpful.

Edit: System details: Linux Mint 18 (~ Ubuntu 16.04) Python 3.5 via Anaconda Haven't faced any issues in any other tasks/exercises in the course so far

[kulshekhar]

Update: I've tried updating anaconda & all the installed packages but I haven't been able to figure out a way to fix this.

Searching for this error, I also came across this stackoverflow question which looks similar but I wasn't able to use the answer to fix this issue

[kulshekhar]

Update: Fixed

For posterity:

This issue is similar to the one posted on this stackoverflow question. The solution was similar. In this case, update the _read32 function in tflearn/dataset/mnist.py from

def _read32(bytestream):
    dt = numpy.dtype(numpy.uint32).newbyteorder('>')
    return numpy.frombuffer(bytestream.read(4), dtype=dt)

to

def _read32(bytestream):
    dt = numpy.dtype(numpy.uint32).newbyteorder('>')
    return numpy.frombuffer(bytestream.read(4), dtype=dt)[0]

I see that the latest source code of tflearn has this fixed. For some reason, installing TFLearn by following the instructions in the lesson wasn't sufficient to avoid this issue.

[arthurtsang]

thanks, that's very helpful. i get the latest by pip install git+https://github.com/tflearn/tflearn.git