Deep Learning in MATLAB - 7) Deep Network Designer

Поделиться
HTML-код
  • Опубликовано: 9 ноя 2024
  • In this video, I go over a cool app that MATLAB has to design and train deep learning networks from scratch.
    #Deep-Learning #MATLAB #AI

Комментарии • 48

  • @anjalir4047
    @anjalir4047 5 месяцев назад

    Wow thank you so much... even after completing matlab onramp course in Deep learning, i am not confident in training my network... now i will start working and write you my results... thank you so much....

  • @mohamedslimmasmoudi6418
    @mohamedslimmasmoudi6418 3 года назад +1

    Very good video Thank You. Masmoudi from Tunisia

  • @Shelvinrw
    @Shelvinrw 3 года назад +1

    Thank you very much. very helpful and detailed in the delivery

  • @hhlol9477
    @hhlol9477 3 года назад +2

    Bro this is AWESOME!!! Thank you very much

  • @ranjansantra
    @ranjansantra 2 года назад

    excellent video! But random=> 1 - so trained network was wrong?

  • @mitikuytc2542
    @mitikuytc2542 Год назад

    It is nice thanks brother

  • @AndrewTran-ig3ri
    @AndrewTran-ig3ri 7 месяцев назад

    thank you❤,the video is awesome!

  • @201120555
    @201120555 Год назад

    a great teacher

  • @gloryprecious1133
    @gloryprecious1133 3 года назад +1

    Useful video thank u. Please upload a video to do semantic segmentation using DeepNeuralDesigner. Thanks in advance.

    • @sunnyguha2
      @sunnyguha2  3 года назад

      Currently I am working on Reinforcement Learning in MATLAB series. Maybe I will take a loop at semantic segmentation after that :)

  • @williamricardobonillavilla1808
    @williamricardobonillavilla1808 3 года назад

    Great tutorial ! Really straigth foward. Thank you for the great contain.

  • @bhanupriyamishra7059
    @bhanupriyamishra7059 2 года назад +2

    Sir if we have exel sheet of features extracted by signals.but that exel sheet is not import in this app. How we import the exel sheet for sequence input data. This app is only working for image nt signal processing?

  • @voom7609
    @voom7609 3 года назад +1

    Can you make a video on how to train two network parallel

    • @sunnyguha2
      @sunnyguha2  3 года назад

      best bet for now : www.mathworks.com/help/deeplearning/ug/deep-learning-with-big-data-on-gpus-and-in-parallel.html

  • @taha546
    @taha546 Год назад

    hello. thanks for the video, i had the same problem you had with sigmoid layer, i removed it and put a relu but the problem is still there, my model is not learning. any suggestions? , also im using big batch of data

  • @farhanahmadbhatti4248
    @farhanahmadbhatti4248 27 дней назад

    in my deep network designer toolbox data and the training options are not showing. I am using Matlab 2024a version. Please guide me about it.

  • @prasad6676
    @prasad6676 2 года назад

    Hi. Can you show me how to classify 3d images of .nii file format. Thanks in advance.

  • @hishan.farfan
    @hishan.farfan 2 года назад

    thank you so much!! 😁

  • @cristhianmanuelduranaceved3748

    Hi, very nice Video and productive. However, I was trying to run the Simulation but I saw this message: Training Failed. Too many input arguments.

  • @jatinprakash4674
    @jatinprakash4674 3 года назад

    Hi, thanks for the valuable information. I'm curious to know about 1D CNN that classifies on the basis of 1D signal input. Can you help me out?

    • @sunnyguha2
      @sunnyguha2  3 года назад

      This might help: www.mathworks.com/help/deeplearning/ug/sequence-to-sequence-classification-using-1-d-convolutions.html

  • @beypazariofficial
    @beypazariofficial 11 месяцев назад

    what up!!!1

  • @muhammadtanveer3194
    @muhammadtanveer3194 3 года назад

    brother do you know how to design RCNN using deep network designer app?

  • @lakshipriyagogoi7858
    @lakshipriyagogoi7858 2 года назад

    How to test this model with other datas.?

  • @محمدمحمدی-ر2د
    @محمدمحمدی-ر2د 2 года назад

    How to enter the matrix in the deep learning toolbox?

  • @ЭдуардШубников
    @ЭдуардШубников Год назад

    How now to check this network?

  • @rudresh7743
    @rudresh7743 3 года назад

    Hey I want to detect traffic cones using raspberry pi camera in matlab. Can you please help me???

  • @TheChristy16
    @TheChristy16 3 года назад

    That was awesome bro, such videos with simple words are what we beginners always expect :) ... can i know, if i can save this net and reuse it later (with same trained weights)? Also, if i reuse that way, is my accuracy expected to improved for the same dataset? please clarify bro.

    • @sunnyguha2
      @sunnyguha2  3 года назад

      Thanks for your kind words. Yes, you can save your network by exporting it to your workspace or saving it using the save() command. You can reuse your network to retrain it to increase accuracy. Just be wary of overfitting :)

  • @manasranjansethi4372
    @manasranjansethi4372 2 года назад +1

    HOW TO GENERATE CONFUSION MATROX FROM THIS

  • @beypazariofficial
    @beypazariofficial 10 месяцев назад

    saa WE LET IT ALL SLIP AWAY

  • @harshavardhanvenkatasaikot6861
    @harshavardhanvenkatasaikot6861 2 года назад

    i got an error regarding the convolution layer 2nd box it shows an error can anyone help me

  • @jayalaxmicsbvc9826
    @jayalaxmicsbvc9826 3 года назад

    how to train this for numerical dataset

    • @sunnyguha2
      @sunnyguha2  3 года назад

      You would have to load in data manually and use featureInputLayer for input

    • @dr.amrgamal3853
      @dr.amrgamal3853 2 года назад

      @@sunnyguha2 how to load data matrix manually?

  • @youssefhakam6807
    @youssefhakam6807 2 года назад

    I don't whu my accuracy is very bad despite of change the sigmoid and add more layers

  • @gdpraveenkumar823
    @gdpraveenkumar823 3 года назад

    Any one Kindly share how to create line chart or bar chart in matlab

  • @muhammadtanveer3194
    @muhammadtanveer3194 3 года назад

    Hi. Anyone help me when I try to export trained network to workspace, it save as series network not DAG network?

    • @sunnyguha2
      @sunnyguha2  3 года назад

      If you network is completely sequential then DAG and series are the same. DAG is used when your network has directed loops (like resnet).

    • @muhammadtanveer3194
      @muhammadtanveer3194 3 года назад

      @@sunnyguha2 thanks brother

    • @muhammadtanveer3194
      @muhammadtanveer3194 3 года назад

      @@sunnyguha2 one thing more, I developed a network it show accuracy about 97% but when i run the same network on the same dataset next time its accuracy decreases to 30%. Do you know what will be the possible reason behind this deceasing trend?

    • @sunnyguha2
      @sunnyguha2  3 года назад

      @@muhammadtanveer3194 seems like you are not saving the trained weights .

    • @muhammadtanveer3194
      @muhammadtanveer3194 3 года назад

      @@sunnyguha2 after training the network i generate code and save the network in workspace. I dnt know why its not working 2nd time?

  • @rolipoli7601
    @rolipoli7601 3 года назад

    why does sigmoid even existed if it's bad?

    • @sunnyguha2
      @sunnyguha2  3 года назад +1

      Sigmoid is sometimes useful as the last layer for logistic decisions and also in shallow and recurrent architectures.

    • @rolipoli7601
      @rolipoli7601 3 года назад

      @@sunnyguha2 thanks!