TensorFlow1.5.0 absl.flags._exceptions.UnparsedFlagAccessError in Jupyter Notebook

[jhljx]

System information

• Have I written custom code (as opposed to using a stock example script provided in TensorFlow): yes
• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Ubuntu 16.04
• TensorFlow installed from (source or binary): binary
• TensorFlow version (use command below): 1.5.0
• Python version: 3.6
• Bazel version (if compiling from source): n/a
• GCC/Compiler version (if compiling from source): n/a
• CUDA/cuDNN version: 9.0/7.0.5
• GPU model and memory:GeForce 950M(4Gb)
• Exact command to reproduce:

Describe the problem

I run the function cifar10.maybe_download_and_extract() of tensorflow/models/tutorials/image/cifar10.py file in python command and jupyter notebook .The code runs successfully in python command, but it causes absl.flags._exceptions.UnparsedFlagAccessErrorfrom cifar10.FLAGS.data_dir in jupyter notebook.

The jupyter notebook has already configured the tensorflow kernel, other tensorflow examples could run smoothly on jupyter notebook.

Significant code in cifar10.py file is:

import tensorflow as tf
import cifar10_input
FLAGS = tf.app.flags.FLAGS

# Basic model parameters.
tf.app.flags.DEFINE_integer('batch_size', 128,
                        """Number of images to process in a batch.""")
tf.app.flags.DEFINE_string('data_dir', '/tmp/cifar10_data',
                       """Path to the CIFAR-10 data directory.""")
tf.app.flags.DEFINE_boolean('use_fp16', False,
                        """Train the model using fp16.""")
# .....
def maybe_download_and_extract():
    """Download and extract the tarball from Alex's website."""
    dest_directory = FLAGS.data_dir
    if not os.path.exists(dest_directory):
        os.makedirs(dest_directory)
# ......

The exception happened from the code statement dest_directory = FLAGS.data_dir.

The successful output in python command is as follows:

>>> cifar10.maybe_download_and_extract()
>> Downloading cifar-10-binary.tar.gz 100.0%
Successfully downloaded cifar-10-binary.tar.gz 170052171 bytes.

The detailed error in jupyter notebook is as follows:

UnrecognizedFlagError                     Traceback (most recent call last)
<ipython-input-6-ddaee26bfd14> in <module>()
----> 1 cifar10.maybe_download_and_extract()
~/models/tutorials/image/cifar10/cifar10.py in maybe_download_and_extract()
381 def maybe_download_and_extract():
382   """Download and extract the tarball from Alex's website."""
--> 383   dest_directory = FLAGS.data_dir
384   if not os.path.exists(dest_directory):
385     os.makedirs(dest_directory)
~/anaconda3/envs/tensorflow/lib/python3.6/site-packages/tensorflow/python/platform/flags.py in 
__getattr__(self, name)
82     # a flag.
83     if not wrapped.is_parsed():
---> 84       wrapped(_sys.argv)
85     return wrapped.__getattr__(name)
86 
~/anaconda3/envs/tensorflow/lib/python3.6/site-packages/absl/flags/_flagvalues.py in __call__(self, argv, known_only)
628       suggestions = _helpers.get_flag_suggestions(name, list(self))
629       raise _exceptions.UnrecognizedFlagError(
--> 630           name, value, suggestions=suggestions)
631 
632     self.mark_as_parsed()
UnrecognizedFlagError: Unknown command line flag 'f'

Source code / logs

#git clone https://github.com/tensorflow/models.git
#cd models/tutorials/image/cifar10
import cifar10
cifar10.maybe_download_and_extract()

[takluyver]

If that function is something you're meant to run from Python code, it shouldn't be looking at sys.argv. If you're meant to run the example as a script, try using %run .../cifar10.py.

[jtpio]

Closing as answered by #3327 (comment), thanks!