Intel NCS Troubleshooting Help and Guidelines

[idata]

Troubleshooting Help and Guidelines

• Looking for documentation on using the NCSDK with your Neural Compute Stick? Please visit https://movidius.github.io/ncsdk/.
• Trying to convert your TensorFlow network for use on the NCS? Please visit https://movidius.github.io/ncsdk/tf_compile_guidance.html.
• Trying to get the NCSDK running on Windows? Please visit https://ncsforum.movidius.com/discussion/comment/1548/#Comment_1548 and see ZhouJustin's comment.
• Interested in the NCSDK and ROS? Please visit https://ncsforum.movidius.com/discussion/403/ros-movidius-ncs-release-v0-4-0/p1 and checkout xhuan28's comment.
• Issues with a VM? Try Oppiie's solution at https://ncsforum.movidius.com/discussion/1239/movidius-solution-to-dispatchereventsend-1-io-error-ncsdk-2-on-vm/p1

Are you having issues? Try the following:

NCSDK Installation Related Checklist

Platform and OS Check

• The NCS was designed to run on a dedicated x86 64 bit machine running Ubuntu 16.04 or a virtual machine system running Ubuntu 16.04. The NCS can also run on a Raspberry Pi running Raspian Stretch and can also run on other x86 based boards including, but not limited to, the UP board and the Minnow board.
• For more information regarding installing the SDK on VMs, please visit: https://movidius.github.io/ncsdk/VirtualMachineConfig.html
• A user has supplied a quick step by step setup with an UP board here: https://ncsforum.movidius.com/discussion/comment/720/#Comment_720

NCSDK Version Check

• You can check the NCSDK change logs to see if a problem existed/was solved in a recent SDK release. This can be done by following these steps:
• Go to https://ncsforum.movidius.com/discussion/98/latest-version-movidius-neural-compute-sdk#latest
• Click on “Latest software release” and scroll down to see the bug fixes and API features for the latest release. We just released NCSDK version 2.08!

Make Examples Error

• Some users have reported receiving make example errors related to previously installed versions of OpenCV and Anaconda. This issue can also be caused by cloning the SDK on a different machine/user and installing it on a Ubuntu machine.
• For OpenCV related install errors, you can visit https://ncsforum.movidius.com/discussion/comment/839/#Comment_839 and try user lengly’s suggestion.
• For users who are using Anaconda, Anaconda seems to be incompatible with the install script. Removing Anaconda has been a solution for two of our users: (https://ncsforum.movidius.com/discussion/comment/639/#Comment_639) ( https://ncsforum.movidius.com/discussion/comment/845/#Comment_845 )
• If the SDK/App Zoo was cloned on a different machine and copied into Ubuntu, you may encounter make example errors. Downloading a fresh copy of the SDK on the Ubuntu machine is recommended and worked for one of our forum users: https://ncsforum.movidius.com/discussion/331/installing-nc-sdk-make-examples-error

Sudo/Root Installation Issues

• Some users have been reporting sudo/root installation errors, this was fixed in the version of the SDK (ver 1.10.00). Please visit https://ncsforum.movidius.com/discussion/98/latest-version-movidius-neural-compute-sdk#latest for more information.

Caffe Import Error

• On the Raspberry Pi, a user had success by reinstalling the python decorator. This was done by typing the following command: sudo apt-get instal -reinstall python*-decorator. Reference post: (https://ncsforum.movidius.com/discussion/comment/1009/#Comment_1009)
• A user reported that installing scikit-image worked for him. This can be done by typing in the following command: sudo pip3 install scikit-image. Reference post: (https://ncsforum.movidius.com/discussion/comment/768/#Comment_768)
• A user reported that when he/she only sees the Caffe Import Error when he/she uses virtualenvs. When the user doesn't use virtualenvs, he did not run into the problem. For more information on virtualenv installation with the NCSDK , please visit https://movidius.github.io/ncsdk/virtualenv.html.
• Make sure that the account that is using the NCSDK is the same account that was used to install the NCSDK.
• Check your bashrc file to make sure your Python path is set.
• Check your Python path by typing in the following command in your Ubuntu terminal to see if your Python path was set by the install script: “echo $PYTHONPATH”
• If your Python path was not set, you can manually set your Python path by adding the following line to your bashrc file: export PYTHONPATH=/opt/movidius/caffe/python:/opt/movidius/mvnc/python:$PYTHONPATH

NCS device error or Timeout error on Raspberry Pi 3

• Using a powered USB hub has helped some community members solve this issue and we recommend using a powered USB hub even if you aren't using a Raspberry Pi. Reference posts: (https://ncsforum.movidius.com/discussion/499/mvncstatus-timeout-raised-after-processing-few-frames-from-video-on-raspberry-pi#latest) and (https://ncsforum.movidius.com/discussion/comment/1437/#Comment_1437)

NCSDK Neural Network Related Checks

Unsupported Networks

• Check with the list of known working networks at:
• Caffe: https://movidius.github.io/ncsdk/Caffe.html
• Tensorflow: https://movidius.github.io/ncsdk/TensorFlow.html

SDK Unsupported layers

• Check the layers (input, hidden, output) used in your network
• Some layers may not be supported by the current NCSDK. View the supported layers here: https://movidius.github.io/ncsdk/Caffe.html. If the layers are not on the supported list, please let us know so that we may accommodate your needs as developers.

>1 Image Inference Not Supported

• Check the batch size of your network’s prototxt file. Currently for the NCS, we only support a batch size of 1.

SDK Layer Errors – Index out of Range

• Make sure to remove all layers related to training and testing. This includes data layers that rely on labeled input.

Graph File Size is too big

• Check the size of the compiled graph file. Compatible graph files should be 320 MB or less.

Caffe Input Shape Construct Error

• In your network’s prototxt file, make sure that the Caffe input shape in the correct format
• Check the correct Caffe input shape specifics by visiting the following link and scrolling down to the Caffe Input layer section: https://github.com/movidius/ncsdk/blob/master/docs/Caffe.md

Trying to run Tiny Yolo V2 on Raspberry Pi?

• Please visit https://ncsforum.movidius.com/discussion/comment/3933/#Comment_3933 on how to run this sample on your Pi.

API related help

• Looking for help with the API? Please visit https://movidius.github.io/ncsdk/ncapi/readme.html.

• Looking to upgrade from NCSDK 1.12 to NCSDK 2.xx? Please visit https://movidius.github.io/ncsdk/ncapi/python_api_migration.html for Python and https://movidius.github.io/ncsdk/ncapi/c_api_migration.html for C/C++.

Still Not Working?

If these suggestions don’t solve your problem(s), please create a forum post with the following information:

1. System specifications (x86? VM? Dedicated machine? Raspberry Pi? Please include OS also)
2. NCSDK version
3. Python version
4. Installation Error Log
5. Miscellaneous Error Log with error message(s)
6. Link to Neural Network for debugging purposes (prototxt file, caffe weights, github, pb files)
7. Any additional relevant information