tony607 / keras-tf-pb

How to convert trained Keras model to a single TensorFlow .pb file and make prediction

How to convert trained Keras model to a single TensorFlow .pb file and make prediction | DLology Blog

How to Run

Require Python 3.5+ and Jupyter notebook installed

Clone or download this repo

git clone https://github.com/Tony607/keras-tf-pb

In the project start a command line run

jupyter notebook

In the opened browser window open

keras_freeze_tf_pb.ipynb

Happy coding! Leave a comment if you have any question.

keras-tf-pb's Issues

TypeError: Keras symbolic inputs/outputs do not implement `op`.

Hi Tony,

Is there any method to solve this problem? I did not find any solution to this problem.

Thank you

ValueError: No model found in config file.

ValueError: No model found in config file.

https://tensorboard.appspot.com/tf-graph-basic.build.html

It appears that the visualization is broken as the link above does not
exists anymore ... what would you suggest for visualization for tf 1.15 ?

Thanks for your support gjk

Show graph throws error on utf-8 encoding on string

Thanks for the demo. I just copy pasted your show graph file into Jupyter using Python 2.7.
It worked just fine. A minor error when using utf-8 encoding on string. Removing encoding='utf-8' in strip_consts function solves it.

ValueError: Cannot create group in read only mode.

---

ValueError Traceback (most recent call last)
in
1 from keras.models import load_model
----> 2 model = load_model('/home/deepedge/mask_rcnn-master/mask_rcnn_damage_0010.h5')

/usr/local/lib/python3.6/dist-packages/keras/engine/saving.py in load_model(filepath, custom_objects, compile)
417 f = h5dict(filepath, 'r')
418 try:
--> 419 model = _deserialize_model(f, custom_objects, compile)
420 finally:
421 if opened_new_file:

/usr/local/lib/python3.6/dist-packages/keras/engine/saving.py in _deserialize_model(f, custom_objects, compile)
219 return obj
220
--> 221 model_config = f['model_config']
222 if model_config is None:
223 raise ValueError('No model found in config.')

/usr/local/lib/python3.6/dist-packages/keras/utils/io_utils.py in getitem(self, attr)
300 else:
301 if self.read_only:
--> 302 raise ValueError('Cannot create group in read only mode.')
303 val = H5Dict(self.data.create_group(attr))
304 return val

ValueError: Cannot create group in read only mode.

Convert to keras-tf-pb with custom loss function

Hi,
I tried you model with a keras model of custom loss function but failed to convert as I'm getting the below error:
FailedPreconditionError: Attempting to use uninitialized value Variable
[[{{node _retval_Variable_0_5}} = _Retval[T=DT_FLOAT, index=5, _device="/job:localhost/replica:0/task:0/device:CPU:0"](Variable)]]

Thanks for your implementation.