Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Sergei Nosov
    @snosov1
    Feel free to ask any questions related to Open Model Zoo!
    neso613
    @neso613
    Hi Sergei

    I am working on OpenVINO for edge deployment of Deep Learning model.
    My task is to deploy model on Raspberrry Pi3 . But my model is very slow. I want to use your Interactive Face Detection model as base. Can I use? If Yes, will you please share python version of Interactive Face Detection demo as C++ code is very hard for me to work with.

    Thanks!!
    Please reply soon.Waiting for your response.

    Sergei Nosov
    @snosov1
    Hey, @nehasoni3 ! Sure, you can use it. We don't really have a Python alternative for interactive face detection, but you can take a look at https://github.com/opencv/open_model_zoo/blob/master/demos/python_demos/object_detection_demo_ssd_async/object_detection_demo_ssd_async.py It's a simpler demo that only calls the detector (no age/gender, landmarks or anything like that). Alternatively, you can also take a look at the (yet, unmerged) Face Recognition demo: opencv/open_model_zoo#151
    Dmitry Kurtaev
    @dkurt

    Hello, Open Model Zoo explorers!

    I'd like to ask you to participate in early stage testing of one of our coming features. This is something called interactive or dynamic OMZ - OpenCV module which let you use every model from the zoo in just few lines of code. Looking forward, it can be also the simplest way to have both C++ and Python implementations of our algorithms and demos without need to port it from one language to another.

    We want to hear from you what do you think about it. The best thing if you can test it and provide any feedback. Let's catch as much bugs as possible :)

    Note that this module can work as with Intel's Inference Engine as without it (for only non-DLDT models). Depends on your machine you can choose one of the instructions from https://github.com/dkurt/open_model_zoo/blob/py_open_model_zoo/ocv_module/open_model_zoo/README.md

    Please leave comments in the corresponding PR: opencv/open_model_zoo#272

    :snake: :wolf: :hear_no_evil: :elephant: :penguin: :octopus: :mouse: :camel: :frog: :ant: :leopard: :fish: :smirk_cat: :beetle: :whale: :koala:

    Daniele@SAS
    @danielecazzari
    Hi Sergei, I have the same C++ to Python issue but for the Human Pose Estimator. The software I need to interface with only has python interface. There is at least some documentation to understand how to handle model input/output the isn't many comments in the code.
    Sergei Nosov
    @snosov1
    Hey, Daniele! Not sure I understand the question fully, but I'll try to be helpful. Firstly, we do lack a Python demo for Human Pose Estimator. One of the reasons is that the demo contains somewhat optimized post processing that was easier to implement in C++ and there's no current plans to create a Python demo. I guess, not so helpful, so far =) Secondly, you can find somewhat more detailed description of the HPE output along with the training code opencv/openvino_training_extensions#83
    TravelToTheTop
    @TravelToTheTop
    Hi everyone, I was wondering if it was possible to port the demos to Android with some tweaks ? Do you think I could make it working thanks to the linux deploy app for example ?
    Sergei Nosov
    @snosov1
    In theory, it should be possible. The open source version of inference engine can be built for ARM and I believe there's even a distribution for Raspberry Pi.
    GotG
    @GotG
    Hello, I can not run myriad_compile in google colab. the error is: can not run binary file. can you please advise? thank you
    Sergei Nosov
    @snosov1
    Hmm, I'm having hard time to understand how and why you want to marry google colab with myriad_compile.
    GotG
    @GotG
    i am using the colab for training and it goes all the way to IR. and it would have been nice to have the IR converted to blob inside the colab. then it can be used directly for myriad x.
    GotG
    @GotG
    so can you please give me any suggestions for what to try ?
    Sergei Nosov
    @snosov1
    sorry, maybe, I'm missing something. If you just call model optimizer to convert - do you still need myriad_compile?
    Dave Malone
    @dave-malone
    Hello, I am attempting to get the open model zoo up and running on a Raspberry Pi 4 with the OpenVINO toolkit. I believe I have everything configured correctly, but when I go to run an inference using the object_detection_sample_ssd with the pre-trained model face-detection-adas-0001, I am getting a very nondescript error that I just can't get past. I'm hoping someone may be able to help. The error I'm getting is [ ERROR ] stoi
    Sergei Nosov
    @snosov1
    @vladimir-dudnik can you, please, advise on the Raspberry Pi support?