when I complie the graph_freeze.pb on the ncs ,it a problem as :

[idata]

from ._conv import register_converters as _register_converters

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

The project graph_freeze.pb is on the web below:

https://github.com/ildoonet/tf-pose-estimation/tree/master/models/graph/mobilenet_thinCan anyone can help me ?

[idata]

It is the Mobilenet on the tensorflow.however ,I can't compile :'(

[idata]

@wangrongxin

请问你找到这个问题的解决方法了吗？请指导，感激不尽！

[idata]

Hi @neilwu

The error you are seeing has been discussed in another thread.

The error is due to the model using variable input sizes which is currently not supported on the NCSDK.

Hope this helps!

Regards,

Jesus

[idata]

Hi @Jesus_at_Intel

Thanks for the information.After I change the model's variable input to a fixed size (432,368), the original error is gone.But I met the problem like https://ncsforum.movidius.com/discussion/comment/2310/#Comment_2310.The errors like this:Fusing depthconv and conv in Openpose/MConv_Stage1_L1_1_depthwise/depthwise and Openpose/MConv_Stage1_L1_1_pointwise/Conv2DTraceback (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 630, in parse_tensor prev_node.addBias( bias.astype(np.float16) )AttributeError: 'int' object has no attribute 'addBias'

If there is any pose estimation demo that can run on Movidius? Please tell me about that.I am going to collapse.

Sincerely,Neil

[idata]

Hi @neilwu

I shared some info regarding the Pose Estimation example using the OpenVINO Toolkit.

https://ncsforum.movidius.com/discussion/1565/pose-estimation-performance#latest

Regards,

Jesus