$30 off During Our Annual Pro Sale. View Details »

Fantastic Models and Where to Find Them

Fantastic Models and Where to Find Them

Rishit Dagli

December 02, 2021
Tweet

More Decks by Rishit Dagli

Other Decks in Technology

Transcript

  1. Fantastic Models
    and Where to
    Find Them
    Rishit Dagli
    High School
    TEDx, TED-Ed Speaker
    rishit_dagli
    Rishit-dagli

    View Slide

  2. Fantastic Models and
    Where to Find Them
    Rishit Dagli

    View Slide

  3. TensorFlow Hub
    rishit_dagli
    A repository of trained
    machine learning models ready
    for fine-tuning and deployable
    anywhere

    View Slide

  4. TensorFlow Hub
    rishit_dagli

    View Slide

  5. ● High School Student
    ● TEDx and 2xTed-Ed Speaker
    ● I contribute to TF Hub
    ● I ❤ ML Research
    ● My coordinates - www.rishit.tech
    $whoami
    rishit_dagli Rishit-dagli

    View Slide

  6. ● Devs who have worked on Deep Learning
    Models (Keras)
    Ideal Audience
    rishit_dagli

    View Slide

  7. rishit_dagli
    ● Off the shelf inference with SoTA
    machine learning models
    Could be super helpful
    rishit_dagli

    View Slide

  8. rishit_dagli

    ● You do need to ask yourself:
    ○ Is there enough information in P to represent Q?
    ○ how much new data?
    ○ can unlabelled data do it?
    ○ what’s the right mix of P and Q data?
    ○ How far is from P from Q?
    Transfer Learning
    rishit_dagli

    View Slide

  9. rishit_dagli

    ● You do need to ask yourself:
    ○ Is there enough information in P to represent Q?
    ○ how much new data?
    ○ can unlabelled data do it?
    ○ what’s the right mix of P and Q data?
    ○ How far is from P from Q?
    ● Samory Kpotufe: Some Recent Insights
    on Transfer-Learning
    Transfer Learning
    rishit_dagli

    View Slide

  10. rishit_dagli

    ● You do need to ask yourself:
    ○ Is there enough information in P to represent Q?
    ○ how much new data?
    ○ can unlabelled data do it?
    ○ what’s the right mix of P and Q data?
    ○ How far is from P from Q?
    ● You can do this easily with TF Hub
    too!
    Transfer Learning
    rishit_dagli

    View Slide

  11. rishit_dagli
    ● Less data
    Could be super helpful
    rishit_dagli

    View Slide

  12. rishit_dagli
    ● Less data
    ● Less domain expertise
    Could be super helpful
    rishit_dagli

    View Slide

  13. rishit_dagli
    ● Less data
    ● Less domain expertise
    ● for different domains including
    image, video, audio etc.
    Could be super helpful
    rishit_dagli

    View Slide

  14. rishit_dagli
    ● Less data
    ● Less domain expertise
    ● for different domains including
    image, video, audio etc.
    ● Take advantage of SoTA machine
    learning models
    Could be super helpful
    rishit_dagli

    View Slide

  15. rishit_dagli
    Ready to Use
    rishit_dagli
    Credits: @TheSandeepGupta

    View Slide

  16. View Slide

  17. Style Transfer

    View Slide

  18. Style Transfer
    rishit_dagli

    View Slide

  19. Style Transfer
    rishit_dagli

    View Slide

  20. Style Transfer
    rishit_dagli

    View Slide

  21. Classification
    Prediction:
    Goldfish,
    Carassius_auratus

    View Slide

  22. Image Classification
    rishit_dagli

    View Slide

  23. rishit_dagli
    bit.ly/hub-classification

    View Slide

  24. Fine tune BERT

    View Slide

  25. Fine tune BERT
    rishit_dagli
    Classify Text with BERT

    View Slide

  26. Fine tune BERT
    rishit_dagli
    Classify Text with BERT

    View Slide

  27. Fine tune BERT
    rishit_dagli
    Classify Text with BERT

    View Slide

  28. Fine tune BERT
    rishit_dagli
    Classify Text with BERT

    View Slide

  29. Fine tune BERT
    rishit_dagli
    Classify Text with BERT

    View Slide

  30. Fine tune BERT
    rishit_dagli
    Classify Text with BERT

    View Slide

  31. Fine tune BERT
    rishit_dagli
    Classify Text with BERT

    View Slide

  32. rishit_dagli
    And more...
    TF Hub for other tasks
    rishit_dagli

    View Slide

  33. rishit_dagli
    tensorflow.org/hub/publish
    Publish your own models

    View Slide

  34. rishit_dagli
    bit.ly/tfhub-deck
    Slides here

    View Slide

  35. Thank You
    rishit_dagli Rishit-dagli

    View Slide