Use This Way Of Training Machine Learning Models For Efficiency

Поделиться
HTML-код
  • Опубликовано: 8 ноя 2024

Комментарии • 76

  • @krishnaik06
    @krishnaik06  2 года назад +5

    Check our courses
    TechNeuron: 200+courses Lifetime Warranty
    Course url: ineuron.ai/one-neuron
    Full Stack Data Scientist Bootcamp 2.0 With Job Guaranteed: ineuron.ai/course/Full-Stack-Data-Science-BootCamp-2.0
    Use Krish20 coupon code for 20% off

    • @varavindhkumar4484
      @varavindhkumar4484 2 года назад +1

      Sir thankyou so much for you and the whole ineuron team for reducing the price of one neuron. I know the offer is 1ly for a week but also u guys have did it sir. I have buyed oneneuron and very happy sir, dont know how to express it. Love you sir.

    • @ganeshsubramanian6217
      @ganeshsubramanian6217 2 года назад

      sir, thank you. Friday - you said it will be NLP live session.. I was waiting for it. When is the next session?

    • @vigkan9610
      @vigkan9610 2 года назад

      Hello @krish Naik the code krish20 is not working for data science bootcamp

    • @SarfrazKhan-yl1un
      @SarfrazKhan-yl1un Месяц назад

      Hey Sir, can you please upload the entire project end to end, which you promised in this video, you've achieved your target of 1K likes on this, and its been 2 years but u haven't posted the project.

  • @AayushOstwal
    @AayushOstwal 2 года назад +24

    Machine Learning is not deterministic, Experiment is the only key!!! Thanks for such content... Kudos!!

  • @yusufhenry9365
    @yusufhenry9365 2 года назад +8

    You are truly a leader!!
    What more can we ask for?
    Thanks my able professor!!

  • @pratikpratik8495
    @pratikpratik8495 2 года назад +8

    Tree algo are not sensitive towards un-scaled features, outlier while LR and other parameteric algo are sensitive towards un scaled feature, outlier. In that case this technique is not useful as you are using the same features for all algo

    • @michaelomosebi6374
      @michaelomosebi6374 2 года назад

      I like this.
      @Krish what do you say?

    • @Nisarg_Nikhil
      @Nisarg_Nikhil 2 года назад

      Correct

    • @sparkersclubesports5743
      @sparkersclubesports5743 2 года назад +2

      @pratik his point here is to use this approach of coding to train multiple models for selection of a good performing model out of that. Ofcourse u can run this code twice for scaled and unscaled data twice with corresponding models. AGREE?

    • @joshuarexmondnunoootoo5316
      @joshuarexmondnunoootoo5316 2 года назад

      Same question I wanted to ask

  • @rishichaitanya3081
    @rishichaitanya3081 2 года назад +2

    For so many days, one thing is running in my mind that like how can we decide the range to take for hyperparameter tuning.

  • @boseashish
    @boseashish 3 месяца назад

    You are doing an amazing job, sir!. God bless you

  • @yousuf1sg
    @yousuf1sg Год назад +5

    Respected Krish Naik Sir, You have to own your committed words in video that you will share the code after 1000 likes. Now more than 1200 and we are still waitig. Thanks

  • @folashadeolaitan6222
    @folashadeolaitan6222 2 года назад +2

    This is amazing! Kudos! Do you mind uploading the notebook please? Thank you

  • @msherazkhan7336
    @msherazkhan7336 2 года назад +3

    Sir... Please upload detail video about tranform models... VIT, POLYVIT, AST... IF AST, its very well becaus my project to use AST Model for sound classification(.wave files)... Thanks

  • @datamonster3212
    @datamonster3212 2 года назад

    I just use LightGBM these days. It usually performs best and requires minimal data preprocessing.

  • @soubhikbandhyopadhyay8160
    @soubhikbandhyopadhyay8160 2 года назад +2

    Krishna sir bohot din se apko follow kar raha hun plz pura video upload kar dijiyega ipy file ke sath plzz....

  • @ankitadeshpande2447
    @ankitadeshpande2447 2 года назад +2

    Kirsh sir please teach also about machine learning model deployment in detail

    • @Eswar.
      @Eswar. 2 года назад +3

      I strongly recommend you to look out campusx (nitish sir) videos on the same
      He was far better

  • @nujhatfatma3476
    @nujhatfatma3476 2 года назад +1

    Sir please share and make a video on this project....

  • @bhabanisankardash7387
    @bhabanisankardash7387 2 года назад +2

    Awesome ....Doing a grt job....hats off

  • @dominickipronoh4324
    @dominickipronoh4324 Год назад

    you gave me a very good idea in this video thanks so much Krish

  • @PalataoArmy
    @PalataoArmy 2 года назад +1

    Would you mind include the eda, feature engineering, better use any boosting model with cv and everything inside pipeline? Thank you sir.

  • @sonubesra771
    @sonubesra771 2 года назад +2

    1000+ likes ✅. Please provide ka whole project implementation

    • @preetjainvd9909
      @preetjainvd9909 2 года назад

      hi , is end to end project upload with ci/cd

    • @ScholarStream_25
      @ScholarStream_25 Год назад

      @@preetjainvd9909 where can I get this code??

  • @dominickipronoh4324
    @dominickipronoh4324 Год назад

    Krish coul you show us how to compare the training and loss accuracies of deep learning models ie to compare new method with existing SOTA methods

  • @ravindarmadishetty736
    @ravindarmadishetty736 7 месяцев назад

    Hi Krish, do you have any plans for Time Series forecasting?

  • @priyabharti-rz3lt
    @priyabharti-rz3lt Год назад

    Sir Please make video on CATBOOST Algorithm

  • @bhuvekkumar4358
    @bhuvekkumar4358 2 года назад +1

    Thank you sir for this amazing code

  • @datasciencegyan5145
    @datasciencegyan5145 2 года назад +1

    @krish 1000 likes completed ... pls upload end to end

    • @preetjainvd9909
      @preetjainvd9909 2 года назад

      Hi , is any update from Krish sir, regard end to end project and code

  • @pratiknaikwade95
    @pratiknaikwade95 2 года назад +1

    Sir plz share this end to end project

    • @preetjainvd9909
      @preetjainvd9909 2 года назад +1

      Are you got reply for end to end implementation with ci / cd, please message.

  • @AI_DOL_
    @AI_DOL_ 2 года назад +3

    Pls sir, is the code available?

  • @sivachaitanya6330
    @sivachaitanya6330 Год назад +1

    sir.......'LogisticRegression' object has no attribute 'accuracy_score' Error how did you get accuracyscore for logistic regresion?

  • @muezahmed7054
    @muezahmed7054 Год назад

    a link for the code would be very appreciated

  • @marcellodichiera
    @marcellodichiera 2 года назад

    Sorry if I am asking something non sense. But instead of doing all of that, why not using AutoML libraries such as Pycaret?

  • @aswiniprakash2567
    @aswiniprakash2567 2 года назад

    Thanks dude, learnt a lot from ur videos 1k crossed pls upload end to end process video

  • @RamsheedMP-h6o
    @RamsheedMP-h6o 5 месяцев назад

    Good video , Thank you @Krish naik

  • @mamunhsn4839
    @mamunhsn4839 2 года назад

    Today Three people came to our Institute for Data Science program. But any how I convinced them not to take admission in our institute .
    I highly suggest them to take admission to iNuron. You are doing a great job Sir👍🏻

    • @krishnaik06
      @krishnaik06  2 года назад

      Which Institute??

    • @mamunhsn4839
      @mamunhsn4839 2 года назад

      @@krishnaik06 EXCELR Raising Excellence Bangaluru

    • @mamunhsn4839
      @mamunhsn4839 2 года назад

      I Wish I Knew It earlier about iNuron.

  • @rohitchan007
    @rohitchan007 Год назад

    Please share the link of the notebook

  • @prashantyadav885
    @prashantyadav885 2 года назад +1

    waiting for this end to end project

  • @mohsinimam2048
    @mohsinimam2048 2 года назад +2

    Sir please upload this project.

  • @AjayBabuR-v2k
    @AjayBabuR-v2k Год назад

    How can I increase the recall?? Even after hyperparameter tuning I was able to increase the recall for class 1 from 8 to 19..

  • @eskayML
    @eskayML 2 года назад +1

    Since trees work generally better than linear algorithms(personal observation) , is there any need of using linear algorithms at all when we are finding best models

    • @johnwig285
      @johnwig285 2 года назад +2

      Ofc, because not everything is about "better" performance i.e. accuracy. It depends on complexity n time n client's's expectations.

  • @bradleymarimbire5603
    @bradleymarimbire5603 2 года назад +1

    Hi sir can we get the code please ?

  • @rangaabi8471
    @rangaabi8471 2 года назад +1

    Can you provide the code sir?

  • @kinjalgogri2850
    @kinjalgogri2850 5 месяцев назад

    it is not showing all only showing Random Forest

  • @AIWALABRO
    @AIWALABRO 2 года назад

    superb..!

  • @rohitsonawane4968
    @rohitsonawane4968 2 года назад

    Sir dsa series continue karo please

  • @rkjyoti4167
    @rkjyoti4167 2 года назад

    superb.

  • @aqib5354
    @aqib5354 2 года назад

    Good afternoon sir

  • @mgaf7864
    @mgaf7864 Год назад

    Sir how to do feature engineer and features selection

  • @ajaykushwaha4233
    @ajaykushwaha4233 2 года назад +3

    Sir one question, instead of trying each algo, can we use automated libraries like pycaret, autoML or other libraries?

    • @HellRaiZOR13
      @HellRaiZOR13 2 года назад

      You can use Lazypredict Library it automatically tests all the available machine learning algorithms in the sklearn Library and give you the score for each of them. Based on that you can an initial decision on which algorithm to start with.
      You dont even have to do the dictionary way like they have shown here.

    • @bruhm0ment767
      @bruhm0ment767 2 года назад

      @@HellRaiZOR13 does it allow parameter tuning?

    • @HellRaiZOR13
      @HellRaiZOR13 2 года назад

      @@bruhm0ment767 no it will train all models on their default settings but later you can pick the best model and then do parameter tuning on those specific models.
      This Library is just to make your model selection process easy and faster.

  • @nashrahkarim7404
    @nashrahkarim7404 11 месяцев назад

    can you provide the code ?

  • @rohitchan007
    @rohitchan007 Год назад

    Please share the code.

  • @sohanaryal
    @sohanaryal 2 года назад

    let's make it it to 1000 likes

  • @engr.FidaHussain
    @engr.FidaHussain Год назад

    too fast, hard to understand

  • @nishantbansal5135
    @nishantbansal5135 2 года назад

    Done

  • @surendhrancbe7872
    @surendhrancbe7872 2 года назад

    Sir..I have a doubt that " in terms of complexity does the project we do in the companies is similar to the project you are doing here ?" since I pursuing degree so I don't have idea about it..that how could be the complexity of the projects in companies. Can anyone please help me out

    • @tykemonster6416
      @tykemonster6416 2 года назад +1

      Not even close. Real world projects are some next level stuff. No work is done in Jupyter, all in visual studio mostly. Also, code is just one part of a project. A big challenge also comes in deployment where you have to dockerize it, create pipelines, then deploy it in cloud platform like Azure/AWS/GCP. Also in deployment another big challenge is some libraries n their versions are supported by some cloud platforms some are not 🤦🏻‍♂️. Then there could be some other technical challenges too in production. It really eats your brain up 😂

  • @pratiknaikwade95
    @pratiknaikwade95 2 года назад +1

    1 like remaining only

  • @Umesh-lz2ro
    @Umesh-lz2ro 6 месяцев назад

    use AutoML do nothing

  • @saravanankarunamurthy9854
    @saravanankarunamurthy9854 2 года назад +2

    Hi Krish, 1000 likes completed.. pls provide GitHub link for this code