iOS example with Image Retraining model always gives the same prediction

[Tugees]

Environment info

Running on: iOS 9.1
Building on OS X El Capitan version 10.11, Xcode version 7.3.1
TensorFlow release 0.9

Steps to reproduce

Follow the contrib/makefile/README to install the tensorflow iOS core lib
Create my own model with the Image Retraining tutorial
Run the iOS example, same error as #2883 is logged.
After fixing errors as suggested same prediction is always given.
Logs or other output that would be helpful

2016-07-21 18:20:06.582 tf_ios_makefile_example[949:323060] Unable to simultaneously satisfy constraints.
    Probably at least one of the constraints in the following list is one you don't want. 
    Try this: 
        (1) look at each constraint and try to figure out which you don't expect; 
        (2) find the code that added the unwanted constraint or constraints and fix it. 
(
    "<NSLayoutConstraint:0x17594000 V:|-(20)-[UIInputSetContainerView:0x176b59d0]   (Names: '|':UITextEffectsWindow:0x17597ab0 )>",
    "<NSLayoutConstraint:0x17593e30 'UIInputWindowController-top' V:|-(0)-[UIInputSetContainerView:0x176b59d0]   (Names: '|':UITextEffectsWindow:0x17597ab0 )>"
)

Will attempt to recover by breaking constraint 
<NSLayoutConstraint:0x17594000 V:|-(20)-[UIInputSetContainerView:0x176b59d0]   (Names: '|':UITextEffectsWindow:0x17597ab0 )>

Make a symbolic breakpoint at UIViewAlertForUnsatisfiableConstraints to catch this in the debugger.
The methods in the UIConstraintBasedLayoutDebugging category on UIView listed in <UIKit/UIView.h> may also be helpful.
2016-07-21 18:20:08.017 tf_ios_makefile_example[949:323060] 1
2016-07-21 18:20:16.855 tf_ios_makefile_example[949:323060] Image path: /var/mobile/Containers/Data/Application/3AEB6ED8-B068-45FA-85D3-36FCFFA21DD1/Documents/photos/image.jpg
I ../tensorflow/contrib/ios_examples/simple/RunModelViewController.mm:302] Session created.
I ../tensorflow/contrib/ios_examples/simple/RunModelViewController.mm:305] Graph created.
I ../tensorflow/contrib/ios_examples/simple/RunModelViewController.mm:310] Creating session.
W tensorflow/core/framework/op_def_util.cc:332] Op BatchNormWithGlobalNormalization is deprecated. It will cease to work in GraphDef version 9. Use tf.nn.batch_normalization().
2016-07-21 18:20:25.542 tf_ios_makefile_example[949:323060] Received memory warning.
2016-07-21 18:20:25.653 tf_ios_makefile_example[949:323060] Received memory warning.
2016-07-21 18:20:38.698 tf_ios_makefile_example[949:323060] Received memory warning.
I ../tensorflow/contrib/ios_examples/simple/RunModelViewController.mm:410] Predictions: 259 4.32 n01342269  moneran, moneron
10 3.69 n00141669   check-in
800 3.67 n01601410  great bowerbird, Chlamydera nuchalis
552 3.59 n01531811  redpoll, Carduelis flammea
0 3.58 n00004475    organism, being

Related to

#2883 I encountered the same error related to DecodeJpeg and then followed the fixes suggested there, but I get the same prediction no matter what I point the camera at. Similarly if i work with the "simple" project and change the photo being identified I still get the same prediction that in the log output above

[petewarden]

I've got an example of how to successfully load and run a retrained model here:
https://petewarden.com/2016/09/27/tensorflow-for-mobile-poets/

I'm going to close this bug, since I believe it's fixed now, please reopen if you're still hitting problems.