incluit / openvino-for-smartcity Goto Github PK

Hi,
I recently bought NCS2 from Intel Movidius Myriad X VPU and I tried this demo on Linux guest (WIN10 host).

Linux guest distribution:
Distributor ID: Ubuntu
Description: Ubuntu 16.04.6 LTS
Release: 16.04
Codename: xenial

All examples woks fine. Now, I want to train another detection model, for example to detect eggs or bulbs or I don't know... something else. It's posible to do that and after could I integreated it into this tutorial?

./intel64/Release/smart_city_tutorial -m intel64/Release/person-vehicle-bike-detection-crossroad-0078/FP32/person-vehicle-bike-detection-crossroad-0078.xml -i ../data/car_1.bmp

output :

Incluit - Openvino-for-Smartcity
CPU
MKLDNNPlugin version ......... 2.1
Build ........... 2020.4.0-359-21e092122f4-releases/2020/4
Could not allocate memory
If missing -d_y argument, try running ../scripts/setupenv.sh

Hi,
When I was running the program using CPU, the performance was outstanding, but when I attempted the run the script

./intel64/Release/smart_city_tutorial -m_vp $vehicle232 -d_vp HDDL -i ../data/NewVideo2.mp4

The performance was poor, with barely 5 frames per second. The device I am currently using for VPU is IEI Mustang-V100-MX8. The device was set up correctly, and it could run the Openvino demos just fine.

Is the program compatible with HDDLs, and has it been tested before? Or is there a problem on my end that I might have overlooked?

Hi,
I recently bought NCS2 from Intel Movidius Myriad X VPU and I follow all steps from this tutorial on Linux guest (WIN10 host).

Linux guest distribution:
Distributor ID: Ubuntu
Description: Ubuntu 16.04.6 LTS
Release: 16.04
Codename: xenial

All steps works fine, even step 4, where installation is testead.

My problem is when I using this script, with MYRIAD parameter (instead, with CPU it works).

./intel64/Release/smart_city_tutorial -m_vp $vehicle232 -d_vp MYRIAD -i ../data/NewVideo2.mp4

This is my console result:

root@ubuntu:/home/redans/tutorials/OpenVino-For-SmartCity/OpenVino-For-SmartCity/build# ./intel64/Release/smart_city_tutorial -m_vp $vehicle232 -d_vp MYRIAD -i ../data/NewVideo2.mp4
Incluit - Openvino-for-Smartcity
[VPU] Unsupported network precision : FP32
If missing -d_y argument, try running ../scripts/setupenv.sh

Do you have any suggestions?

Hi,
In Ubuntu x64, v. 16_04, I run successfully this demo and worked perfectly, using CPU. Now I want to detect another thinghs like coins or bulbs or ..... something else, not just cars or pedestrian. I don't know how to create and assing a new trained model to this demo. Any suggestions? What should I use for this TensorFlow, Caffe or Yolo alg. ?

Hi,

I am trying to test the project and I get the following error with OpenVino R5 on an Ubuntu 16.04.
What can be?

/usr/bin/ld: CMakeFiles/smart_city_tutorial.dir/src/object_detection.cpp.o: referencia sin definir al símbolo '_ZN15InferenceEngine10TensorDescC1ERKNS_9PrecisionESt6vectorImSaImEENS_6LayoutE'
/opt/intel/computer_vision_sdk_2018.5.445/deployment_tools/inference_engine/lib/ubuntu_16.04/intel64/libinference_engine.so: error adding symbols: DSO missing from command line
collect2: error: ld returned 1 exit status
CMakeFiles/smart_city_tutorial.dir/build.make:288: fallo en las instrucciones para el objetivo 'intel64/Release/smart_city_tutorial'
make[2]: *** [intel64/Release/smart_city_tutorial] Error 1
CMakeFiles/Makefile2:69: fallo en las instrucciones para el objetivo 'CMakeFiles/smart_city_tutorial.dir/all'
make[1]: *** [CMakeFiles/smart_city_tutorial.dir/all] Error 2
Makefile:127: fallo en las instrucciones para el objetivo 'all'
make: *** [all] Error 2

Hi!
Would it be possible to use this example to track pedestrians with multiple cameras?
If not, some idea of what should be changed?
Thank you,

I got this error when trying to run the first tutorial and I ran successfully the demo of openvino in the documentation, any help will be grateful
and thank you in advance^^

Hello.

I want to know what is the latest version of OpenVINO that thse demos can work (or confirmed work) and branched to be used.

So far, I could successfull run SmartCity and Driver-Behavior demos, using "OpenVINO<=2019R1" branch of code, with OpenVINO R1.0.1.

This is because your README said follows:

https://github.com/incluit/OpenVino-For-SmartCity/blob/master/README.adoc

The latest OpenVINO™ toolkit installed and verified. Supported versions +2018 R4.0. (Lastest version supported 2019 R1.0.1)

However, 2019 R1.0.1 (API version 1) is kind of obsolete now and installation package is no longer available from Intel website.

Also, if you try using newer OpenVINO versions, I get build errors due to OpenVINO API differnces (they
are chaning API for newer versions so demo apps written for version 1 APIs are no longer buildable with
newer OpenVINO.)

So, I wonder if there is any chance the demos were ported to newer version of OpenVINO and if so, what is the instruction for build.

Thank you your your comments or advice!