MultiModal Search (Text+Image) using TF MobileNet , HF SBERT in Python on Kaggle Shopee Dataset

Поделиться
HTML-код
  • Опубликовано: 22 дек 2024

Комментарии • 22

  • @DeepDiveinUniverse
    @DeepDiveinUniverse 9 месяцев назад

    Can we use it in some kind of recommendation system's ?

    • @RitheshSreenivasan
      @RitheshSreenivasan  9 месяцев назад

      This was just a multimodal similarity match for search

  • @anilchaudhry804
    @anilchaudhry804 3 года назад

    Amazing video sir. You are helping alot of students and aspiring early professional data scientists like us. Please keep up the good work and keep posting content. Just a small suggestion, can you solve live problems on kaggle? We would like to learn the thought process while solving. If so it would be very helpful

    • @RitheshSreenivasan
      @RitheshSreenivasan  3 года назад

      Thank you. Kaggle is tough and needs a lot of time. So I try to look at Kaggle for datasets and ideas. Then try out something offline

    • @anilchaudhry804
      @anilchaudhry804 3 года назад

      Sir how can one contact you?

    • @RitheshSreenivasan
      @RitheshSreenivasan  3 года назад

      Connect with me on linkedin

    • @anilchaudhry804
      @anilchaudhry804 3 года назад

      @@RitheshSreenivasan sure sir

  • @navinbondade5365
    @navinbondade5365 3 года назад

    cant we train neural network for distinguishing between similar items, we could create a random dataset for model to learn non similarities or we use conditional modelling

    • @RitheshSreenivasan
      @RitheshSreenivasan  3 года назад

      You can train neural networks called Siamese network to distinguish similar items

  • @venkatesanr9455
    @venkatesanr9455 3 года назад

    Thanks for the informative video, sir. I believe we can use gensim cosine similarity functions here? if I am wrong kindly correct me

  • @navinbondade5365
    @navinbondade5365 3 года назад

    amazing video really like your content

  • @tejpalkumawat776
    @tejpalkumawat776 3 года назад

    Nice project sir .
    Can we use cosine similarity for our task , I mean it can also find the top 5 nearest neighbour from the training data?

    • @RitheshSreenivasan
      @RitheshSreenivasan  3 года назад +1

      I have used cosine similarity only. That’s why I have to normalise the embedding vector

  • @alipaloda9571
    @alipaloda9571 3 года назад

    thanks for such video its pretty informative however i am not able to open your git can you help in that
    thanks

    • @RitheshSreenivasan
      @RitheshSreenivasan  3 года назад

      GitHub should be accessible to anyone

    • @alipaloda9571
      @alipaloda9571 3 года назад

      @@RitheshSreenivasan I know sir can you check that I am talking about that particular file (multi modal one)

    • @RitheshSreenivasan
      @RitheshSreenivasan  3 года назад

      I am able to view the repository without logging in to GitHub so not sure what’s the issue

  • @swathys7818
    @swathys7818 2 года назад

    Hi Sir, intresting problem statement and excellent code walk-through. Sir, Could you please share your suggestion on using Zero shot learning or contrastive learning for similarity detection module. Your suggestion and inputs will be very helpful for us Sir

    • @RitheshSreenivasan
      @RitheshSreenivasan  2 года назад +1

      Contrastive learning can be used for similarity detection.

    • @swathys7818
      @swathys7818 2 года назад

      @@RitheshSreenivasan thank you sir