What is Self Attention | Transformers Part 2 | CampusX

Поделиться
HTML-код
  • Опубликовано: 18 дек 2024

Комментарии •

  • @balrajprajesh6473
    @balrajprajesh6473 10 месяцев назад +148

    A̶t̶t̶e̶n̶t̶i̶o̶n̶ Nitish Sir is all you need.

  • @adityamehra1414
    @adityamehra1414 10 месяцев назад +28

    Sir you are at very close to breakout. ML/DL in India will take wind fast soon. These days everything changes fast . Your vision of becoming the hub for this content , i see it very clearly. I am forever indebted to you , as you have helped me save a lot of time and anxiety in this information overloaded times , where quality is rare to find.

  • @aj_ai
    @aj_ai 10 месяцев назад +5

    Kudos to the Nitish Sir!! who made self-attention a simple melody.

  • @ali75988
    @ali75988 10 месяцев назад +2

    For people like me, this man is the coach, brother, & guide. Thank you so much Nitish. Loads of prayers and love.

  • @ali75988
    @ali75988 10 месяцев назад +2

    Thankyou so much Nitish, whatever deep learning I know is because of you and might be able to switch my career with this learning.

  • @ankanmazumdar5000
    @ankanmazumdar5000 10 месяцев назад +5

    I am an NLP engineer, but no one I have ever seen to explain in such simpler manner...God bless you Nitish

    • @ganeshvhatkar9040
      @ganeshvhatkar9040 7 месяцев назад

      how do you get into this role, sir? I'm doing a Master's in AI from a tier-1 college, Please suggest me a path.

  • @sachinbagale563
    @sachinbagale563 6 месяцев назад +6

    From past 3 months , I am only giving "ATTENTION" to your videos. Mad Respect!

  • @myself4024
    @myself4024 4 месяца назад +5

    🎯 Key points for quick navigation:
    00:00 *🎥 Introduction to Self Attention*
    - Introduction by Nitesh, explaining the importance of understanding self-attention for mastering transformers,
    - Announcement of the video being part of a series, and the structure of the series.
    01:18 *📚 Importance of Self Attention in NLP*
    - Explains how understanding self-attention helps in mastering transformers,
    - Mentions that transformers are crucial for advanced AI applications like LLMs and generative AI.
    02:14 *❓ Key Requirement in NLP Applications*
    - Highlights the need to convert words into numbers for any NLP application,
    - Discusses various NLP tasks like sentiment analysis and machine translation.
    03:30 *🔢 Vectorization Techniques*
    - Introduces vectorization as the process of converting words to numbers,
    - Talks about basic techniques like one-hot encoding and its limitations,
    - Describes the initial steps in NLP focusing on efficient word-to-number conversion.
    05:46 *🧳 Bag of Words*
    - Explains the Bag of Words technique as an improvement over one-hot encoding,
    - Describes how Bag of Words considers word frequency in sentences.
    06:40 *🔍 TF-IDF and Word Embeddings*
    - Discusses the limitations of Bag of Words and introduces TF-IDF,
    - Introduction to word embeddings as an advanced technique,
    - Emphasizes the importance of capturing semantic meaning with word embeddings.
    08:14 *🌐 Creating Word Embeddings*
    - Explanation of the process of creating word embeddings using large training datasets,
    - Describes how neural networks are used to convert words into n-dimensional vectors,
    - Examples of how similar words have similar vectors in multi-dimensional space.
    10:18 *🔍 Dimensionality and Meaning*
    - Discusses how each dimension in a word embedding can represent different aspects of word meaning,
    - Explains how similar words have similar vectors due to shared contextual meanings,
    - Mentions the inherent problem of neural networks not explicitly defining dimension meanings.
    11:39 *⚠️ Limitations of Word Embeddings*
    - Describes a potential issue with word embeddings when the training data focuses heavily on a specific word,
    - Uses a hypothetical example to explain how over-representation of certain words can skew embeddings,
    - Explains the importance of balanced training data to avoid biased embeddings.
    13:30 *🍎 Word Embedding Averaging Issue*
    - Describes the averaging issue with word embeddings,
    - Example of how different contexts for the word "apple" affect its embedding,
    - Emphasizes that embeddings capture average meaning, not specific contextual meaning.
    17:02 *🛠️ Static Nature of Word Embeddings*
    - Explains the static nature of word embeddings and their limitations,
    - Discusses how static embeddings can cause problems in applications like translation,
    - Highlights the need for contextual embeddings that adjust based on sentence context.
    19:05 *✨ Introduction to Self-Attention*
    - Introduces self-attention as a solution to the limitations of static word embeddings,
    - Describes self-attention's role in generating smart contextual embeddings,
    - Explains how self-attention adjusts embedding values based on context dynamically.
    21:01 *📚 Mechanism of Self-Attention*
    - Describes how self-attention generates contextual embeddings,
    - Each input word embedding produces a contextual output embedding,
    - These embeddings understand the context of each word, making them more useful for NLP applications like transformers.
    22:09 *🧩 Self-Attention Summary and Next Steps*
    - Summarizes self-attention as a mechanism for generating contextual embeddings from static embeddings,
    - Highlights the improved utility of these embeddings for NLP applications,
    - Introduces the next video’s focus on the calculations inside self-attention, including concepts like query, key, and value vectors.
    Made with HARPA AI

  • @MALIKXGAMING804
    @MALIKXGAMING804 10 месяцев назад +1

    Amazing explanation of Attention mechanism, Bag of words, and embedding. Let me say superb you are real guru

  • @BadBoy-yb5pq
    @BadBoy-yb5pq 10 месяцев назад +2

    Sir I dont Know how to Thank you , I am a student from kerala who really intrested in the field of Machine Learning and AI. I can easly say how lucky Iam to find your class , I was completely a beginner in this field and i found out your 100 days of ML , It was my turning point, you made my base undersatnding Very well ,Now I can almost do all the ML work flow by my self but iam not good at it though. I need More practices ,I tried almost all project on your Channel and It made me More Under Standing about the things and Once More THANKSSS. now as i said i can do all ml work flow but iam not much good at any of and i know it need more Practices and Mainly some Domain Knowelge in each field . Sir Can you give Some advice to increase my progress like how should i practice now ,should I repeate the practice (like taking some competition dataset and then do all the work then check others work , compare to yourself), and Lastly and Mainly I would appreciate if you do More ML and Data analysis Project on your Channel so I can Increase My domain Knowledge and diffrent feature engineering technique (Sorry about my broken english)

  • @educationalpoint725
    @educationalpoint725 10 месяцев назад +2

    I was eagerly waiting for the next video for last 1 week.
    Thank you for providing this amazing content.
    Also, I request you to please release the other videos of transformer soon.

  • @mudassir_fayaz
    @mudassir_fayaz 6 месяцев назад +2

    Such an amazing explanation.
    You have also the talent of self-attention,
    Based on the student context you have explained very well.

  • @shinwarikhan4677
    @shinwarikhan4677 10 месяцев назад +1

    thank you so much sir for such an amazing concepts, sir please make video how it work when there is images not textual data

  • @rahulgaikwad5058
    @rahulgaikwad5058 10 месяцев назад +1

    Amazing video on what is self attention waiting for upcoming videos, thank you so much for all your hardwork and efforts 😊

  • @smitgandhi3865
    @smitgandhi3865 Месяц назад

    Thank you sir for the wonderful overview of self-attention

  • @drrabiairfan993
    @drrabiairfan993 10 месяцев назад +1

    You are a guru!

  • @TusharKumar-ly4mk
    @TusharKumar-ly4mk 10 месяцев назад

    Nitish Sir is one-stop solution for all the AI-related problems

  • @tarunkr4698
    @tarunkr4698 10 месяцев назад +1

    The Best video on Internet!!!

  • @yamansaini6379
    @yamansaini6379 5 месяцев назад

    Khatarnak Explanation 👏🏻👏🏻👏🏻

  • @gauravmalik3911
    @gauravmalik3911 10 месяцев назад

    as always, explanation is detailed , in-depth, intuitive, looking forward for further videos

  • @mohammadarif8057
    @mohammadarif8057 10 месяцев назад +1

    God Bless You ..Master

  • @prodigy4836
    @prodigy4836 10 месяцев назад

    Thank you so much for keeping such quality content free!
    Please do not stop making these videos.

  • @manishsingh-cb3pp
    @manishsingh-cb3pp 10 месяцев назад

    This is Ultimate ....you have made the embedding easy and created attention i.e very nicely explained

  • @DineshKumar-cu3bg
    @DineshKumar-cu3bg 5 месяцев назад

    You are superb in making everyone learn and understand. Hats off to you

  • @sanskaragrawal3526
    @sanskaragrawal3526 10 месяцев назад

    Thankyou For this Video @CampusX, I am learning all of the concepts of deep learning from your playlist so I was waiting for this video, Thank you for this great playlist of Deep learning. this playlist is very helpful for me in understanding Deep Learning.

  • @SambitSatapathy-e4b
    @SambitSatapathy-e4b 10 месяцев назад

    So easily explained. Thanks and heartfelt gratitude for such quality content. 🙏🙏🙏🙏

  • @KumR
    @KumR 10 месяцев назад

    Thank you so much Nitish. I know this was a very tough nut to break it into simple words with examples. Looking forward to the next set of videos which could solve the "how" mystery and how it also flows into multiheaded attention. Once again Kudos for the excellent effort and for great service to community.

  • @GauravKumar-gh6qj
    @GauravKumar-gh6qj Месяц назад

    Great and very simple explanation

  • @shivombhargava2166
    @shivombhargava2166 10 месяцев назад

    Never thought such a complicated topic could have a simple one-liner explanation.

  • @PrashantParhad
    @PrashantParhad 10 месяцев назад

    Nice explanation Sir. I am very excited for next 2-3 videos...

  • @vidhikumar1664
    @vidhikumar1664 9 месяцев назад

    Hi Nitish,
    Thank you so much for these intuitive detailed explanations!

  • @NabidAlam360
    @NabidAlam360 10 месяцев назад +1

    Please upload the other parts soon too!

  • @Riyanpatel-g2y
    @Riyanpatel-g2y 2 месяца назад

    Such an amazing explanation.

  • @curiousverse7363
    @curiousverse7363 4 месяца назад

    The way you explain is amazing!!!

  • @maryamirshad8492
    @maryamirshad8492 10 месяцев назад +1

    bro you nailed it but i need next videos too, i have interview tomorrow🥲. Anyways love from PAKISTAN

  • @you_goat_code
    @you_goat_code 3 месяца назад

    Sir ap ki video buhut lambi hoti hai magar Jo knowledge milta hai wo unbelievable hai. Thanks for dedicating your important time for us.
    Your videos are late but they worth it❤❤❤.

  • @priyankatomar-k2e
    @priyankatomar-k2e 10 месяцев назад

    Great content...keep going... waiting for next parts 😊

  • @narendraparmar1631
    @narendraparmar1631 4 месяца назад

    Very well explained , Thanks

  • @amriteshnandan4681
    @amriteshnandan4681 10 месяцев назад +1

    sir, please make one video on encoder model like BERT after you finish this series.

  • @ghousepasha4172
    @ghousepasha4172 5 месяцев назад

    Love the way u explain topics ❤❤❤❤❤

  • @tarun80534
    @tarun80534 3 дня назад

    Bhai great level of explanation

  • @charanpoojary4804
    @charanpoojary4804 Месяц назад

    kya batau sir????? badiya bahut badiya

  • @rameshwarsingh5859
    @rameshwarsingh5859 10 месяцев назад

    NITISH YOU ARE THE BEST MENTOR EVER

    • @rameshwarsingh5859
      @rameshwarsingh5859 10 месяцев назад

      And please nitish grow bro With new Unique content like how to read and interpret any new DL papers ...in want u have the family of millions subscribers

  • @kamaljeetsingh8701
    @kamaljeetsingh8701 10 месяцев назад

    Thankyou so much for creating such a content. Waiting for next videos related to transformers.

  • @zahidullah1707
    @zahidullah1707 8 месяцев назад

    Sir really appreciate your kind efforts. I have a humble Suggestion, please organize this series properly so that the beginner levels can follow it accordingly.

  • @rajnishadhikari9280
    @rajnishadhikari9280 9 месяцев назад

    very nice lecture sir
    needed part 3 and part 4 of this transformers series

  • @NabidAlam360
    @NabidAlam360 10 месяцев назад

    Thank you so much for your video! You are special brother!

  • @nirjharjajodia7169
    @nirjharjajodia7169 8 месяцев назад

    Next level teaching...

  • @sarmadafzalj
    @sarmadafzalj Месяц назад

    Amazing explanation!

  • @ParthivShah
    @ParthivShah 7 месяцев назад +1

    Thank You Sir. Great Explanation.

  • @imteyazahmad9616
    @imteyazahmad9616 10 месяцев назад

    Amazing ❤
    Waiting for next

  • @tanishagrawal9091
    @tanishagrawal9091 7 месяцев назад

    You deserve lot more sir🙏

  • @AbdurRahman-x1p5l
    @AbdurRahman-x1p5l 23 дня назад

    Best Explanation!!!

  • @Ahm77887
    @Ahm77887 10 месяцев назад

    Thank you very much for your efforts sir😍❤....

  • @sagarkale2252
    @sagarkale2252 9 месяцев назад +1

    Nitish sir you are need❤

  • @agaming4187
    @agaming4187 2 месяца назад

    great teacher

  • @surajhulke6196
    @surajhulke6196 10 месяцев назад

    Always excited 😊

  • @ariousvinx
    @ariousvinx 10 месяцев назад

    Thank you sir for this series ❤

  • @bigbossfan007
    @bigbossfan007 Месяц назад

    beautifully explained

  • @sagarbhagwani7193
    @sagarbhagwani7193 10 месяцев назад

    Again Thankusomuch sir❤

  • @haz5248
    @haz5248 10 месяцев назад

    Been waiting for this 🔥 . . . Ik what im gonna do tonight .

  • @sachink9102
    @sachink9102 6 месяцев назад

    Awesome !!!!!
    NitishJi Kaha the ab tak aap .

  • @shinwarikhan4677
    @shinwarikhan4677 10 месяцев назад +1

    sir how it work in case of images please make video on it

  • @rockykumarverma980
    @rockykumarverma980 18 дней назад

    Thank you so much sir 🙏🙏🙏

  • @Amanullah-wy3ur
    @Amanullah-wy3ur 5 месяцев назад

    this is helpful 🖤🤗

  • @kashifsadiq4145
    @kashifsadiq4145 10 месяцев назад

    great Sir

  • @bloodline39
    @bloodline39 5 месяцев назад

    What about differentiating all English words based on parts of speech? Like we normal humans do l, and then training a model to understand how the parts of speeches for together to form a sentence?

  • @emadshamim2754
    @emadshamim2754 10 месяцев назад

    sir, kindly please upload the remaining videos on architecture of the transformers

  • @shreeganeshak6346
    @shreeganeshak6346 10 месяцев назад

    Thank you sir 🙏

  • @unbking4190
    @unbking4190 3 месяца назад +1

    sir if embedding gives values based on average meaning so how do embedding knows in which context is words used in data as machine can only understand numbers and by seeing numbers how can it conclude context

  • @user-km9sm8pj8c
    @user-km9sm8pj8c 10 месяцев назад +1

    sir make a video on ViT

  • @kamilshaikh1602
    @kamilshaikh1602 8 месяцев назад

    amazing!

  • @defaulter0715
    @defaulter0715 10 месяцев назад +3

    kis kis ko lagata hai ki nitish sir krish naik se THE BEST hai....??

    • @ComicKumar
      @ComicKumar 10 месяцев назад +1

      dono alag alag hai bhai...krish sir first step hai to understand a topic...and nitish sir poora depth mein samjhate hain!

  • @aiforeveryone
    @aiforeveryone 10 месяцев назад

    Amazing

  • @lnsexplorer2128
    @lnsexplorer2128 10 месяцев назад

    Thank you sir

  • @muhammadikram375
    @muhammadikram375 10 месяцев назад

    sir kindly also continue Mlops playlist 😢❤

  • @cr_sehgal
    @cr_sehgal 10 месяцев назад

    Sir, great video as always
    how can we contact you?
    i wanted to discuss a project-related topic with you

  • @akshayshelke5779
    @akshayshelke5779 9 месяцев назад +1

    Waiting for new videos on this topic.

  • @sibaprasadnaikbehera3442
    @sibaprasadnaikbehera3442 10 месяцев назад

    love your explanation.how to translate a new language ?

  • @eyeonmystery
    @eyeonmystery 10 месяцев назад

    Yes sir next vedio plz

  • @hari_1357
    @hari_1357 10 месяцев назад

    Thank you

  • @durgeshameta254
    @durgeshameta254 5 месяцев назад

    ur genius

  • @RakibRohan-ph6zj
    @RakibRohan-ph6zj 2 месяца назад

    Please complete this Deep Learning Playlist

  • @rayyt6364
    @rayyt6364 8 месяцев назад

    [Just about anything] -> [numbers] -> Neural Network. This is literally entire ML

  • @technicalhouse9820
    @technicalhouse9820 10 месяцев назад +1

    Sir 💖♥♥💚💚💛💛💜💜

  • @khushalrathi6280
    @khushalrathi6280 10 месяцев назад

    sir please jaldi se ye playlist complete kar di jiye...please sir

  • @bhaveshkumarchoubisa9228
    @bhaveshkumarchoubisa9228 4 месяца назад

    love u sir

  • @kmishy
    @kmishy 9 дней назад

    In this beginning era of teaching RNN, DL, Transformer etc, how are you the best! May I know where you learned

  • @ShubhamChaudhari-wj2gc
    @ShubhamChaudhari-wj2gc 10 месяцев назад

    Sir can you please upload all the videos ASAP 😢😢

  • @AshishSharma-tf3fy
    @AshishSharma-tf3fy 6 месяцев назад

    sir pls complete this playlist , we need llms

  • @kanakkanak1520
    @kanakkanak1520 10 месяцев назад

    when is the next batch of DSMP starting?

  • @haz5248
    @haz5248 10 месяцев назад

    quick quick baki ke videos bhi dal do

  • @pxxthik
    @pxxthik 10 месяцев назад

    💖💕

  • @not_amanullah
    @not_amanullah 10 месяцев назад

    Understood++

  • @znyd.
    @znyd. 6 месяцев назад

  • @DataTalesByMuskan
    @DataTalesByMuskan Месяц назад

    what if the sentence was Apple launched an iphone while i was eating apple .So will there be 2 different embeddings for "apple" ?

  • @AsphaltNitro-py3hg
    @AsphaltNitro-py3hg 10 месяцев назад

    Sir videos roj upload kariye na

  • @UPSCTHINKER
    @UPSCTHINKER 10 месяцев назад

    Data

  • @aadityagupta5776
    @aadityagupta5776 9 месяцев назад

    *is*

  • @hemantkp_
    @hemantkp_ 9 месяцев назад +1

    What an amazing explanation sir!! The way you explain concepts from scratch is just phenomenal. 🫡🫡