New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Bug]openvino2022 compile_tool failed. Failed to allocate graph: MYRIAD device is not opened. #11807
Comments
Have you installed a pre-built version of OpenVINO (contains the MyriadX-device plugin) or have you compiled it from source-code (doesn't contain the plugin)? |
I just want to convert the IR model into a blob. I don't want to run the model in the real-world here. |
@brmarkus I have the same problem (tried in openvino docker container and also in openvino compiled from source code). |
@changewOw @Luxonis-Vlad Thank you for reporting, I am able to reproduce the issue on 2022.1 release. The compile tool should not require the Neural Compute Stick 2 to be connected. I confirmed the compile tool works as expected in 2021.4.2 release and I reported this issue to the development team. Regards, Ref. 86400 |
having same issue here as well |
Any update @jgespino ? |
@szabi-luxonis Not yet, I will link the pull request once we have a fix. |
Same issue here. I just want to convert the IR model into a blob. |
Hi all! This can be worked around by using the following config option: As soon as there is a proper fix, I will let you know. |
Thanks @DariaMityagina for the workaround, looking forward to the proper fix. @alexk-ede @szabi-luxonis @kamegbor @changewOw @Luxonis-Vlad Can you try the workaround provided by @DariaMityagina ? It worked for me using the following steps.
|
Indeed, it worked. |
Hi, I am facing the same issue in the docker container running on ubuntu 20.04. `openvino@4582fb921620:/opt/intel/openvino_2022.2.0.7713$ /bin/bash -c "omz_downloader --name googlenet-v1 --precisions FP16 && omz_converter --name googlenet-v1 --precision FP16 && curl -O https://storage.openvinotoolkit.org/data/test_data/images/car_1.bmp && python3 samples/python/hello_classification/hello_classification.py public/googlenet-v1/FP16/googlenet-v1.xml car_1.bmp MYRIAD" ========== Downloading /opt/intel/openvino_2022.2.0.7713/public/googlenet-v1/googlenet-v1.prototxt ========== Downloading /opt/intel/openvino_2022.2.0.7713/public/googlenet-v1/googlenet-v1.caffemodel ========== Replacing text in /opt/intel/openvino_2022.2.0.7713/public/googlenet-v1/googlenet-v1.prototxt ========== Converting googlenet-v1 to IR (FP16) Model Optimizer arguments: Please install required versions of components or run pip installation % Total % Received % Xferd Average Speed Time Time Time Current |
@Veereshdammur this looks like a different topic; when only compiling an IR-model to a blob (e.g. compiling a model on a HOST-machine, then deploying the compiled model to a production environment) it should not be required to have a Myriad-X/NCS2 device connected. Your topic sounds like MyriadX was not set-up properly on your HOST-machine (like UDEV-rule not installed) and/or the way you started your Docker container is missing USB-related information and/or you have created your own Docker container but missed the usb-udev-event-related information (see the note under "https://docs.openvino.ai/latest/openvino_docs_install_guides_installing_openvino_docker_linux.html#known-limitations"). Which OS are you running on, have you installed OpenVINO on the HOST, i.e. outside the container? Can you use a pre-built OpenVINO Docker container (see "https://hub.docker.com/u/openvino") and then try the command line as shown under "https://docs.openvino.ai/latest/openvino_docs_install_guides_installing_openvino_docker_linux.html#running-the-image-on-intel-neural-compute-stick-2"? |
Hi, Thanks for the quick response. I am actually trying to configure the NCS2 on my ubuntu machine. So far I have been attempting to install openvino-dev via PIP but couldn't succeed. I also made an attempt to run the docker image in a container as specified in the documentation (link). The setup details are mentioned below, OS version: Ubuntu 20.04.5 LTS Docker version: The NCS 2 device is plugged into a USB3.0 slot and it's getting detected within the system :
I have pulled the ubuntu20_dev image from the remote hub:
And then I issued this command from the documentation of NCS2, but it was failed due to invalid argument.
Then I tried issuing the second command (uses the privilege mode) and it worked.
When I tried running the same example which is given in the documentation (link), I was encountered with the error
So, it seems like Udev rules are not updated while running the command. How do we solve this problem? |
Hi!
|
Thanks a lot for your answer. It worked!! |
System information (version)
Detailed description
When I use the compile_tool for conversion, I meet an error.
[setupvars.sh] OpenVINO environment initialized
/home/zyb/2661
OpenVINO Runtime version ......... 2022.1.0
Build ........... 2022.1.0-7019-cdb9bec7210-releases/2022/1
Network inputs:
image : u8 / [N,C,H,W]
Network outputs:
detection_out/sink_port_0 : f16 / [...]
Failed to allocate graph: MYRIAD device is not opened.
Steps to reproduce
omz_downloader --name face-detection-0200 -o ./omz/
source setupvars.sh
./compile_tool -m ./omz/intel/face-detection-0200/FP32/face-detection-0200.xml -d MYRIAD -ip U8
Issue submission checklist
The text was updated successfully, but these errors were encountered: