Product of input blob dimensions C,H,W (96,3448,1) must be equal to the parameter 'inputChannels' (96)

[Mahanotrahul]

I have converted a tensorflow frozen model using tfcoreml converter to mlmodel format. However, while loading the model for prediction
model = MLModel('file.mlmodel')
the error shown is this -

RuntimeWarning: You will not be able to run predict() on this Core ML model. 
Underlying exception message was: Error compiling model: "compiler error:  
Inner product layer: 'decoder/MatMul_34' : Product of input blob 
dimensions C,H,W (96,3448,1) must be equal to the parameter 'inputChannels' (96)

I am using Python2.7 (outside venv)
coremltools 3.0b1
MacOS 10.14.4

I am unable to understand the problem here.
In python3.7, the model loads successfully without any error. However, there calling predict on the model generates this issue #317 #372.

Exception: Unable to load CoreML.framework. Cannot make predictions.

How do I solve this issue?
Thanks!

[aseemw]

Seems like the model is invalid. Can you share the model? You can also file a bug report and attach the model to it.

[slin07]

In the tfcoreml conversion did you add use_coreml_3=True?

import tfcoreml as tf_converter
tf_converter.convert(tf_model_path = 'my_model.pb',
                     mlmodel_path = 'my_model.mlmodel',
                     output_feature_names = ['softmax'],
                     input_name_shape_dict = {'input' : [1, 227, 227, 3]},
                     use_coreml_3 = True)