Add DBpedia classification example

[ilblackdragon]

DBPedia classification like in "Character-level Convolutional Networks for Text Classification" is a good example to showcase:

• Character level inputs
• CNN for characters
• show how much simpler than https://github.com/zhangxiangxiao/Crepe

[ilblackdragon]

Done. The models in examples are a lot simpler and work on word and character level RNN and CNN.

[marziehsaeidi]

Hi, I am trying to run Text Classification examples but it seems datasets is not recognised under skflow:

module 'skflow' has no attribute 'datasets'

[terrytangyuan]

You need to install the most recent release of TensorFlow and check the examples there. Please post issues to TensorFlow repo.

[zd123]

Hey, if you dont want to re-install the nightly build, you can manually download the same dataset. I've documented ez steps here: http://stackoverflow.com/questions/37206459/tensorflow-examples-all-fail-due-to-attributeerror-module-object-has-no-attri

[jeffheaton]

I think some compile errors have been introduced to this example from all the TF changes

[neuralminds]

thanks for these examples. just tried with Text Classification for character_rnn now how can i write a serving_input_fn for it ?
extended the code to save but getting error, please help
from tensorflow.contrib.learn.python.learn.utils import input_fn_utils
feature_spec = {"feature":tf.FixedLenFeature([100],tf.int64)}
serving_input_fn = input_fn_utils.build_parsing_serving_input_fn(feature_spec)
and than
classifier.export_savedmodel(export_dir_base='model', serving_input_receiver_fn=serving_input_fn)

and getting this error

TypeError: Failed to convert object of type <class 'dict'> to Tensor. Contents: {'feature': <tf.Tensor 'ParseExample/ParseExample:0' shape=(?, 100) dtype=int64>}. Consider casting elements to a supported type.

Thanks