LSTM: Understanding the Number of Parameters

แชร์
ฝัง
  • เผยแพร่เมื่อ 12 ม.ค. 2025

ความคิดเห็น • 42

  • @diegoperezromero6610
    @diegoperezromero6610 ปีที่แล้ว +1

    Sir your explanation is the best I have found on TH-cam by far. Thank you so much for sharing it helped a lot!

    • @MuratKarakayaAkademi
      @MuratKarakayaAkademi  ปีที่แล้ว

      Thank you so much for sharing your opinion! Please keep learning and commenting :)

  • @boxiongtan69
    @boxiongtan69 2 ปีที่แล้ว +1

    The best explanation for LSTM structure (8:54) on TH-cam

  • @ronakpandya6086
    @ronakpandya6086 2 ปีที่แล้ว +4

    Thank you sir for a clear explanation of LSTM parameter calculations. It will be very useful for visualizing the weights and the internal structure of LSTM.

  • @raviprakash8014
    @raviprakash8014 2 ปีที่แล้ว +1

    Thanks a lot for this video . I learned about LSTM after reading few books and some articles . i had some 'loose' idea and many questions and but could not find answer of my doubts any where . Your video not only helped to validate my understanding but answered many other queries that i had . So grateful that you shared this knowledge.

    • @MuratKarakayaAkademi
      @MuratKarakayaAkademi  2 ปีที่แล้ว

      Glad to hear that! Please keep learning and writing comments 🙏 😊

  • @Deshammanideep
    @Deshammanideep 3 ปีที่แล้ว +1

    Struggling to understand LSTM from a long time. This video is eye opener.

  • @pablosrn442
    @pablosrn442 3 ปีที่แล้ว +1

    finally someone who explain it properly, thanks !

  • @RadomName3457
    @RadomName3457 2 ปีที่แล้ว

    Thanks a lot sir, you cleared my confusion regarding this topic

  • @englishwitharash
    @englishwitharash 2 ปีที่แล้ว +1

    You did an awesome job explaining this. Thanks 🙏

  • @ngogioan2411
    @ngogioan2411 23 วันที่ผ่านมา +1

    Thank you for your excellent sharing. I've got a little bit confusion, that's when you create LSTM layer, sometimes you use output = LSTM(LSTMoutputDimension)(input) and sometimes you use output = LSTM(numberOfLSTMcells)(input). What is the correct understanding? Because in LSTM: How it works? How to use? How to set up parameters correctly? video, you mention "units parameter" as numberOfLSTMcells but in this video, you mention "units parameter" as outputDimention

    • @MuratKarakayaAkademi
      @MuratKarakayaAkademi  9 วันที่ผ่านมา

      Thank you for the question. You are right. The correct term must be dimension however in many resources mistakenly it is referred as cells. Actually there is no cell or unit in the lstm structure as.

  • @techshark7194
    @techshark7194 3 ปีที่แล้ว +1

    Amazing content....thanks a ton for your effort to make such videos.

  • @habtelejebo895
    @habtelejebo895 2 ปีที่แล้ว +1

    Excellent, never seen before

  • @bijoygayen2641
    @bijoygayen2641 3 ปีที่แล้ว +1

    Thank you very much, sir, for dedicating such an eye-opening video to understanding the LSTM parameters. Sir, I have a question At 19:25 why did you multiply each W, U,b parameter with 4 (line number 6,8,10)??? The second question is in every time step (i.e., 5 timesteps) h and c state will change??? Basically, I want to reproduce the LSTM using NumPy by Keras derived LSTM parameters, so what parameters should I keep on focus??

    • @MuratKarakayaAkademi
      @MuratKarakayaAkademi  3 ปีที่แล้ว +1

      Thank you for the questions. 1. We multiply each W, U, b parameter with 4 because there are 4 gates in an LSTM cell. 2. Each time step, the LSTM cell creates new h and c states. 3. The input and the output parameters define the internal behavior of the LSTM cell.

    • @bijoygayen2641
      @bijoygayen2641 3 ปีที่แล้ว +1

      @@MuratKarakayaAkademi Thank you sir...

  • @keren718
    @keren718 ปีที่แล้ว +1

    Teacher I have a question, the hidden state and cell state dimension is 2 but the input dimension is 3, does that mean when we concatenate them we have to pad? Thank you very much for this tutorial

    • @MuratKarakayaAkademi
      @MuratKarakayaAkademi  ปีที่แล้ว

      Hi @keren718, great question! When the hidden state and cell state dimensions are 2, and the input dimension is 3, you'll need to ensure that the dimensions match when you concatenate them. In this case, you should consider reshaping or padding the hidden state and cell state to match the input dimension of 3. This will allow for proper concatenation. Thanks for watching the tutorial, and don't hesitate to ask if you have more questions! 😊

  • @arshiagharagozlou9768
    @arshiagharagozlou9768 8 หลายเดือนก่อน

    Thank You!

    • @MuratKarakayaAkademi
      @MuratKarakayaAkademi  8 หลายเดือนก่อน

      You're welcome! Please keep learning and commenting 👍 🙏

  • @NiranjanBallal
    @NiranjanBallal 3 ปีที่แล้ว +1

    I have a question. If I am giving a one-to-many approach.
    Let's say I have 6 input parameters, then my input vector would be (batch, 1, 6) and
    if I want to predict an output parameter value for 500 time steps, my output vector would be (1, 500).
    Now, should my LSTMoutputDimension be 500 or 1? (if I have return sequence = True).

    • @MuratKarakayaAkademi
      @MuratKarakayaAkademi  3 ปีที่แล้ว

      First, remember that LSTM expects input data to be a 3D tensor such that: [batch_size, timesteps, feature] Please note that timesteps and features are two different concepts! If you want to predict an output parameter value for 500 time steps, than the shape would be something like [batch_size, 500, feature]! Not [batch_size, timesteps, 500]! THese are two different outputs! PLease first understand the problem and the expected output in terms of [timesteps, feature]!I hope ıt is clear now.

  • @vinayKumar-ot3ic
    @vinayKumar-ot3ic 3 ปีที่แล้ว +1

    Excellent

  • @yasirabdulkareem9844
    @yasirabdulkareem9844 2 ปีที่แล้ว

    💯

  • @akshitchaturvedi2775
    @akshitchaturvedi2775 7 หลายเดือนก่อน +1

    Remove the background music, it's distracting.

    • @MuratKarakayaAkademi
      @MuratKarakayaAkademi  7 หลายเดือนก่อน +1

      THank you for the feedback. I will try it. Take care!

  • @ShahzebKhan-pd3yj
    @ShahzebKhan-pd3yj 6 หลายเดือนก่อน +1

    remove the music from videos

    • @MuratKarakayaAkademi
      @MuratKarakayaAkademi  6 หลายเดือนก่อน

      Thank you for the feedback. Unfortunately, in TH-cam studio, I can't remove the music. The music ends after 01:40. Take care!

  • @habtelejebo895
    @habtelejebo895 2 ปีที่แล้ว +1

    Excellent, never seen before

    • @MuratKarakayaAkademi
      @MuratKarakayaAkademi  2 ปีที่แล้ว +1

      Thank you!

    • @habtelejebo895
      @habtelejebo895 2 ปีที่แล้ว +1

      @@MuratKarakayaAkademi please one more video on training lstm model using particle swarm optimization. Thanks

    • @MuratKarakayaAkademi
      @MuratKarakayaAkademi  2 ปีที่แล้ว +1

      I noted your request.