Build a Validation Set With TensorFlow's Keras API

Поделиться
HTML-код
  • Опубликовано: 25 янв 2025

Комментарии • 29

  • @deeplizard
    @deeplizard  4 года назад +2

    👉 Check out the blog post and other resources for this video:
    🔗 deeplizard.com/learn/video/dzoh8cfnvnI
    👀 Come say hey to us on OUR VLOG:
    🔗 ruclips.net/user/deeplizardvlog

  • @tickleu9838
    @tickleu9838 4 года назад +12

    Just found this channel 3 days ago. Have already gone through the machine learning fundamentals and now started with the implementation playlist. The incremental approach of teaching is excellent and the playlist is extremely well ordered. Great job on the effort you guys put (even including the link to the corresponding videos whenever you mention a concept). Keep it up. Gained a subscriber and I am looking forward to the upcoming videos as well

  • @tymothylim6550
    @tymothylim6550 3 года назад +2

    Thank you very much for this video! It really helped me to see how setting aside a validation set can be done this way, as well as the need to shuffle the data beforehand!

  • @zerocel
    @zerocel 4 года назад +5

    NICE!!! so useful!!! thanks

  • @Sikuq
    @Sikuq 4 года назад +1

    I have enjoyed this video/blog several times to fully understand the shuffling/validation on-the-fly complex. Great playlist item. Thanks.

  • @navsiv11
    @navsiv11 4 года назад +2

    hi, mary, I'm a huge fan of you both, you ppl helped me a lot in understanding various concepts in AI, im so happy to see you both finally!!!

  • @shreekanthajith6783
    @shreekanthajith6783 4 года назад +2

    Hi :) , I appreciate your work, it's really helpful. Keep doing what you are doing . Thanks.

  • @shalinisgoel6881
    @shalinisgoel6881 2 года назад +1

    hi, your teaching skills are great...and most importantly an easy explanation of difficult topics makes it more excellent....thanks for all the efforts..
    how can i access all 43 videos??
    I can see only 21 videos here and also on the blog post there are 21 videos and text....

    • @deeplizard
      @deeplizard  2 года назад +1

      You are most welcome, Shalini. The second half of the original TensorFlow course has now been separated out into its own course regarding neural network deployment linked below.
      deeplizard.com/learn/playlist/PLZbbT5o_s2xrwRnXk_yCPtnqqo4_u2YGL2

    • @shalinisgoel6881
      @shalinisgoel6881 2 года назад

      @@deeplizard alright....thanks a ton n again ur videos are best😍

  • @Jtube0101Mega
    @Jtube0101Mega 3 года назад

    Great lesson!

  • @debatradas9268
    @debatradas9268 3 года назад +1

    thank you so much

  • @sarvatmir5888
    @sarvatmir5888 4 года назад +1

    Amazing content .Thanku

  • @LayneSadler
    @LayneSadler 4 года назад

    I've just been doing 2 scikitlearn splits. I forget why but a lot of stack exchange people were complaining about how validation_data was implemented

  • @Mikuwl
    @Mikuwl 4 года назад

    Do you guys have any videos on distributional reinforcement learning?

  • @navsiv11
    @navsiv11 4 года назад +1

    please start video series on BERT, XLNET and other NLP SOTA models please

  • @benjaminlim1735
    @benjaminlim1735 4 года назад

    HI, i am a little confused on the validation set. Do we need to manually create a validation set from the training set by manually removing say 10% of the training data and putting in into a validation dataset before training begins. Or does keras automatically does it for us after declaration a validation_split in model.fit ? Or that both has to be done before training begins.

    • @deeplizard
      @deeplizard  4 года назад

      You can do it either way. In this episode, we focused on the latter way of doing it (Keras splitting the validation set out for us). In the corresponding blog, there is further discussion for how to do it the first way (manually creating a validation set prior to training).
      deeplizard.com/learn/video/dzoh8cfnvnI

    • @benjaminlim1735
      @benjaminlim1735 4 года назад +1

      @@deeplizard Yes, thanks. I saw it. I really enjoy watching your video and have greatly benefited from it and thanks so much for sharing all these knowledge.
      From Singapore.

  • @navsiv11
    @navsiv11 4 года назад

    can you also start computer vision playlist having GANS and Transformers

  • @gosee7914
    @gosee7914 4 года назад +1

    I love you so much

  • @richarda1630
    @richarda1630 3 года назад

    sorry nerd question: what is the hexa after "History at.." in the output? Is that a memory address? if so , just curious as to why Keras would include such low-level info.

    • @deeplizard
      @deeplizard  3 года назад

      Ha, I've never looked into it :D

  • @thebrothershow5826
    @thebrothershow5826 4 года назад +2

    WOW

  • @ScriptureFirst
    @ScriptureFirst 3 года назад

    kFold? Maybe in a future vid?

  • @sazzakursaju5806
    @sazzakursaju5806 4 года назад

    just four in joy

  • @brandonlee9528
    @brandonlee9528 3 года назад

    She is very very attractive

  • @tostupidforname
    @tostupidforname 4 года назад

    So uhm whenever i run the model it basically starts at 97% accuracy. Why is that? My guess was that its is because model is still defined in the jupyter notebook but i restarted kernel and it still happens.