Community
cancel
Showing results for 
Search instead for 
Did you mean: 
idata
Community Manager
369 Views

Questions about following Guidance for Compiling TensorFlow Networks

Hi there,

 

I've been using NCSDK to follow the steps "Guidance for Compiling TensorFlow Networks" on the website of Intel Movidius.

 

Successed in the step creating three files : mnist_model.index, mnist_model.data-00000-of-00001, mnist_model.meta,

 

but when I tried to modifiy mnist_deep.py into mnist_deep_inference.py and continued running the code, the following two line is the message I got:

 

2019-05-01 02:06:15.543110: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA

 

2019-05-01 02:06:15.543955: I tensorflow/core/common_runtime/process_util.cc:69] Creating new thread pool with default inter op setting: 2. Tune using inter_op_parallelism_threads for best performance.

 

The first Question appears: I'm not sure what it tries to tell me and what should I do.

 

In order that whether I failed or not , so I checked the file. To my surprise, it does generate: mnist_inference.index, mnist_inference.data-00000-of-00001, mnist_inference.meta..

 

So I typed the command "mvNCCompile", but it tells me "mvNCCompile: command not found". :'(

 

The second question is why mvNCCompile doesn't work since I'm using NCSDK?

 

Really need your help,please. :'(

Tags (1)
0 Kudos
1 Reply
idata
Community Manager
80 Views

Hi @yi-ru

 

Please try to re-install the Neural Compute SDK as I mentioned in your other thread:

 

https://ncsforum.movidius.com/discussion/1724/mvnccompile-not-found#latest

 

Could you also answer the following question about your setup?

 

     

  • Which OS are you running on your computer?
  •  

  • Are you running a virtual machine?
  •  

  • How much memory does your system have?
  •  

  • What processor does your system have?
  •  

 

Regards,

 

Jesus
Reply