DL Workbench on Edge Infer Error

[faridzin]

User ID - u74471

JOB ID = 162790

When inferring, at 86% got an error that says "Got unrecognized code error from the server".

Using DL Workbench on my local machine gives this error :

[ ERROR ] Check 'seq_mask_pshape[0] == time_size' failed at core/src/op/ctc_greedy_decoder.cpp:84:

What does this mean? Is it a problem with my model?

[ChrisB_Intel · Accepted solution]

faridzin,

There appears to be a problem with that model. I ran it and got the same error, so we are looking at that now. In the mean time do you need to run that specific model, or were you just trying out DevCloud? If you need that model we were able to convert license-plate-recognition-barrier-0007 and run it. Additionally, while it is a different model, for a different purpose, we also ran vehicle-license-plate-detection-barrier-0106.

Not sure if any of that helps you in the short term while we look at that model?

Thanks

Chris

[artyom_tugaryov]

Hello @faridzin.,

In the screenshot that you provided, I noticed that the "precision" field on the table is empty. I guess that there is a problem with loading the model to the CPU device. So as I understand, we not able to infer this model on CPU.

Can you say, how you upload this model? from OMZ tab or from your machine?

Thanks,

Artyom 

[faridzin]

Hello @artyom_tugaryov ,

I uploaded this model from my machine. I can confirm it was a problem with my model.

Thank you and @ChrisB_Intel  for your help!