Exception in the Custom model example

[mhjabreel]

Hi,

I have tried running the custom model example and i got the following exception

Traceback (most recent call last):
File "", line 1, in
File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/base.py", line 160, in fit
monitors=monitors)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/estimator.py", line 449, in _train_model
train_op, loss_op = self._get_train_ops(features, targets)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/estimator.py", line 673, in _get_train_ops
_, loss, train_op = self._call_model_fn(features, targets, ModeKeys.TRAIN)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/estimator.py", line 656, in _call_model_fn
features, targets, mode=mode)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/base.py", line 369, in _model_fn
predictions, loss = model_fn(features, targets)
File "", line 3, in my_model
TypeError: dnn() got an unexpected keyword argument 'keep_prob'

Any ideas what might be causing this?

[ilblackdragon]

Which example did you run? Please use examples from https://github.com/tensorflow/tensorflow/tree/master/tensorflow/examples/skflow
Also if something doesn't work, please post in that repo (this repo is inactive at this point).

[ilblackdragon]

https://github.com/tensorflow/tensorflow/blob/master/tensorflow/examples/skflow/iris_custom_model.py shows how to use new tf.contrib.layers for dnn with dropout.