Intel® Distribution of OpenVINO™ Toolkit
Community assistance about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all aspects of computer vision-related on Intel® platforms.
6392 Discussions

Error while converting tensor flow model

idata
Employee
897 Views

Hi,

 

I am trying to convert faster_rnn_inception_v2_coco_2018_01_28 model found at tensorflow model zoo using mvnccompile and i get following error:

 

user@ubuntu:~/movidius/faster_rcnn_inception_v2_coco_2018_01_28$ mvNCCompile frozen_inference_graph.pb -in image_tensor -on detection_scores -s 12 -is 600 600 -o coco.graph

 

mvNCCompile v02.00, Copyright @ Movidius Ltd 2016

 

/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/ops.py:766: DeprecationWarning: builtin type EagerTensor has no module attribute

 

EagerTensor = c_api.TFE_Py_InitEagerTensor(_EagerTensorBase)

 

/usr/local/lib/python3.5/dist-packages/tensorflow/python/util/tf_inspect.py:45: DeprecationWarning: inspect.getargspec() is deprecated, use inspect.signature() instead

 

if d.decorator_argspec is not None), _inspect.getargspec(target))

 

/usr/local/lib/python3.5/dist-packages/h5py/init.py:34: FutureWarning: Conversion of the second argument of issubdtype from float to np.floating is deprecated. In future, it will be treated as np.float64 == np.dtype(float).type.

 

from ._conv import register_converters as _register_converters

 

/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/tensor_util.py:509: DeprecationWarning: The binary mode of fromstring is deprecated, as it behaves surprisingly on unicode inputs. Use frombuffer instead

 

return np.fromstring(tensor.tensor_content, dtype=dtype).reshape(shape)

 

Traceback (most recent call last):

 

File "/usr/local/bin/mvNCCompile", line 118, in

 

create_graph(args.network, args.inputnode, args.outputnode, args.outfile, args.nshaves, args.inputsize, args.weights)

 

File "/usr/local/bin/mvNCCompile", line 104, in create_graph

 

net = parse_tensor(args, myriad_config)

 

File "/usr/local/bin/ncsdk/Controllers/TensorFlowParser.py", line 259, in parse_tensor

 

input_data = np.random.uniform(0, 1, shape)

 

File "mtrand.pyx", line 1307, in mtrand.RandomState.uniform

 

File "mtrand.pyx", line 242, in mtrand.cont2_array_sc

 

TypeError: 'NoneType' object cannot be interpreted as an integer

 

Can someone let me know what am i missing?

 

The link for the model: http://download.tensorflow.org/models/object_detection/faster_rcnn_inception_v2_coco_2018_01_28.tar.gz

0 Kudos
4 Replies
idata
Employee
474 Views

I have a project using Mask RCNN so I tried to make sure it can run with NCS. But after searching on the forum, I found NCS has not support RCNN yet.

 

Here's the thread: https://ncsforum.movidius.com/discussion/765/mask-rcnn

 

Maybe it's the reason you cannot convert your model.

0 Kudos
idata
Employee
474 Views

@appati Here is the release note in which they list all the supported/tested networks:

 

https://movidius.github.io/ncsdk/release_notes.html
0 Kudos
idata
Employee
474 Views

@apatti Thanks Hiankun for posting a link to all of the supported networks. Currently we don't have support for faster r-cnns. Regarding object detectors, we do support SSD Mobilenet and Tiny Yolo v1 on Caffe and Tiny Yolo v2 on Tensorflow.

0 Kudos
idata
Employee
474 Views

@Tome_at_Intel Thank you very much for your answers in the forums - I was reading through them these days. We are interested in using Movidius USB stick, with Python3, Ubuntu 16.04 and networks trained with Python3 TensorFlow's Object Detection API: https://github.com/tensorflow/models/tree/master/research/object_detection

 

Do you have plans to support the RCNN models (or others) from that API?

 

We are facing the same problem of not being able to compile a trained network for your Movidius stick,

 

using either the ones supported in Model Zoo: https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md)

 

or custom trained using pretrained models from that zoo.

 

Thank you very much for your time!

0 Kudos
Reply