Using ML.NET for Object Detection in WinForms

Поділитися
Вставка
  • Опубліковано 21 гру 2024

КОМЕНТАРІ • 32

  • @AJ93-lx6on
    @AJ93-lx6on 7 місяців тому

    I tried to train dataset in google colab with yolov5 and downloaded it in onnx format model still it did not work. Can you create tutorial for yolov5 google colab model training and then export onnx model to predict in winform c#.

  • @gornat5839
    @gornat5839 4 роки тому +7

    Great tutorial, but can you tell me where is MLModel folder in git with onnx file?

    • @JonWood
      @JonWood  3 роки тому +1

      I didn't include it since I got the files from Custom Vision which was with this video - ua-cam.com/video/zr6M7Lzr48w/v-deo.html

  • @zoltan.halasz
    @zoltan.halasz 4 роки тому +1

    I discovered your channel from your traversy media appearance. Great job 🙂

  • @riclin9531
    @riclin9531 2 роки тому

    sorry I have a problem
    I can only detect the picture that I have labeled it
    is that correct?
    because when I try a picture that I haven't labeled it.
    it can be detected

  • @DougSchofield
    @DougSchofield 3 роки тому +1

    Everything was going well until 4:10 when he did something from "custom vision". I went to the link he specified to find thoses files but they didn't exist. This was a waste of time.

    • @JonWood
      @JonWood  3 роки тому +1

      The files are from the Microsoft Custom Vision service which I go into in this video - ua-cam.com/video/zr6M7Lzr48w/v-deo.html

  • @dangnguyenvan5500
    @dangnguyenvan5500 2 роки тому

    Dear sir.
    i got an error like this: System.InvalidOperationException: 'Column data doesn't match input node names of model.'
    could you tell me how to fix this error,
    hope you feedback!
    Thanks in advance!!!

  • @rakeshnarayanan3584
    @rakeshnarayanan3584 2 роки тому

    Can we fetch similar names from database using ML.Net using this algorithm example Name Entity Relationship ?

  • @manuelchristianmalig-on5114
    @manuelchristianmalig-on5114 4 роки тому +1

    hello, is it possible to detect live feed on the picturebox?

    • @JonWood
      @JonWood  4 роки тому +1

      Like from a video capture? I’m working on something like that, actually

    • @projectartemis8218
      @projectartemis8218 4 роки тому

      @@JonWood hello, why does the end of my ONNX exported file is string instead of a float?

    • @semdejong9849
      @semdejong9849 3 роки тому +1

      @@JonWood could I ask you a few questions about that?

    • @JonWood
      @JonWood  3 роки тому +1

      @@semdejong9849 Ask away! :)

  • @fmr2921
    @fmr2921 2 роки тому

    Hi jon Wood. In video you user model ONNX have input data (Nonex3x416x416) but i user model ONNX have input images (16x3x1024x1024) when run is error.
    Can you help me fix error !

  • @h7cking
    @h7cking 3 роки тому

    How to call yolox model

  • @ankursoni8060
    @ankursoni8060 3 роки тому

    How can we get the X and Y co ordinate of the detected image ?

  • @RGUKT_714
    @RGUKT_714 Рік тому

    Please make a video on Object detection and Image classification using Custom vision in Web application or Blazor application apart of console application

    • @JonWood
      @JonWood  Рік тому

      I plan to redo the custom vision with object detection video since some new changes have come out since then. I'll work on using a Blazor app to make the predictions for it.

  • @canergurelsoy6875
    @canergurelsoy6875 4 роки тому +1

    Very good and advanced content,
    And for newbies, do you also have this same content for REST API as WinForms can run in a Linux environment?
    and what if we need to show the boundingBoxes in more than one place in a single image?
    Thanks for sharing.

  • @viktoralferov2874
    @viktoralferov2874 2 роки тому

    It works (at home)! Thank you so much. But sometime "chardonnay" detect like "red" ) I need more different images for training.

  • @spekbalazs6304
    @spekbalazs6304 Рік тому

    Hi,
    I tried copying the code but when i run it i get an error message that says: System.ArgumentOutOfRangeException: 'Could not determine an IDataView type and registered custom types for member Image
    Paraméter neve: rawType'. Do you have any idea what could be wrong?

    • @JonWood
      @JonWood  Рік тому +1

      Just off the top of my head it looks to be something with the Image property. I'm also pretty sure this may have changed since this video came out. I plan on making a brand new one that goes end-to-end with Custom Vision and object detection in ML.NET if that would help.

  • @gametechtutorial4670
    @gametechtutorial4670 2 роки тому

    Nice sharing knowledge bro😁👍

  • @hopexstudio6388
    @hopexstudio6388 3 роки тому

    Doesn't work on x86?

    • @JonWood
      @JonWood  3 роки тому

      ML.NET is x64 only, so that wouldn't work on x86.

  • @agvasparagos
    @agvasparagos 4 роки тому +1

    As a novice ML developer, I'm not able to learn from this video.
    There errors are handled fast and without explanations, and code in the Git is full of build/compile errors, so there is no reference that I can correct my errors from.

    • @JonWood
      @JonWood  4 роки тому

      What compile errors are you getting and which ones didn’t get explained well?

    • @agvasparagos
      @agvasparagos 4 роки тому

      @@JonWood
      Thanks for the reply.
      5:27 The errors are in line 12.
      You fix it by adding "Using System.Drawing;
      I didn't catch that before...
      Also,
      In 3:54 you add two folders - MLModel and Models
      After Creating the folders, you copy 2 files into MLModel.
      The 1st one is just a text file - lebels.txt, with Red & White.
      The 2nd is model.onnx, and it's not clear where to get it from and what it is doing.
      In 6:36 you mention a video, can you add a link to it?
      I assume there will be some explenation on the netron app.
      This also comes back to the file model.onnx, that is missing.
      I'll keep trying to get this code working on my machine :)
      Thanks!