Part 1-Road To Learn Finetuning LLM With Custom Data-Quantization,LoRA,QLoRA Indepth Intuition

Поделиться
HTML-код
  • Опубликовано: 28 ноя 2024
  • Quantization is a common technique used to reduce the model size, though it can sometimes result in reduced accuracy.
    Quantization-aware training is a method that allows practitioners to apply quantization techniques without sacrificing accuracy. It is done in the model training process rather than after the fact. The model size can typically be reduced by two to four times, and sometimes even more.
    Fine Tuning Playlist: • Steps By Step Tutorial...
    -------------------------------------------------------------------------------------------------
    Support me by joining membership so that I can upload these kind of videos
    / @krishnaik06
    -----------------------------------------------------------------------------------
    ►AWS Bedrock Playlist: • Generative AI In AWS-A...
    ►Llamindex Playlist: • Announcing LlamaIndex ...
    ►Google Gemini Playlist: • Google Is On Another L...
    ►Langchain Playlist: • Amazing Langchain Seri...
    ►Data Science Projects:
    • Now you Can Crack Any ...
    ►Learn In One Tutorials
    Statistics in 6 hours: • Complete Statistics Fo...
    End To End RAG LLM APP Using LlamaIndex And OpenAI- Indexing And Querying Multiple Pdf's
    Machine Learning In 6 Hours: • Complete Machine Learn...
    Deep Learning 5 hours : • Deep Learning Indepth ...
    ►Learn In a Week Playlist
    Statistics: • Live Day 1- Introducti...
    Machine Learning : • Announcing 7 Days Live...
    Deep Learning: • 5 Days Live Deep Learn...
    NLP : • Announcing NLP Live co...
    ---------------------------------------------------------------------------------------------------
    My Recording Gear
    Laptop: amzn.to/4886inY
    Office Desk : amzn.to/48nAWcO
    Camera: amzn.to/3vcEIHS
    Writing Pad:amzn.to/3OuXq41
    Monitor: amzn.to/3vcEIHS
    Audio Accessories: amzn.to/48nbgxD
    Audio Mic: amzn.to/48nbgxD

Комментарии • 56

  • @riteshsingh811
    @riteshsingh811 9 месяцев назад +15

    Your DS-ML videos helped me to crack Big4 3 years ago & again after so many years I am coming back to my favourite teacher to learn about GenAI. Understanding these complicated subjects and then teaching it to us requires another level of dedication . Thanks a lot Krish for constant hardwork. You are a boon to all the students out there.

  • @kirthankiran
    @kirthankiran 9 месяцев назад +16

    Founder of ineuron and still consistent on RUclips and teaching 😮. Great

  • @RahulKumar-lv9yz
    @RahulKumar-lv9yz 9 месяцев назад +14

    Yesterday only I requested this video, and here it is today.
    Thankyou so so much Krish!! Can't express enough gratitude.

  • @Abdullah_kwl
    @Abdullah_kwl 9 месяцев назад +6

    Thanks for explaining quantization, please also explain more topics such as PEFT LoRA QLoRA AI3 MoE, you are the only person who can explains these topics with intuition and mathematical concepts. Keep it up. ❤

  • @hamidraza1584
    @hamidraza1584 9 месяцев назад +2

    Lots of love and respect from Lahore Pakistan.i have very clear understanding of artificial intelligence by watching your RUclips videos

  • @stat_life
    @stat_life 28 дней назад

    I am an GenAI Engineer and this is the most in-depth video on quantization i had ever seen.

  • @arupmmi07
    @arupmmi07 9 месяцев назад

    Thank you very much for your hard work, I don't know what I'll do if you don't post this video. I think your the only creator in this entire RUclips community who described this topic so well. I am waiting for your LoRA, & QLoRA video.
    Thanks.

  • @Arbajsofficial
    @Arbajsofficial 9 месяцев назад +4

    You are really awesome Krish sir , my most of the skills are gift from you.... But still not getting job need help

  • @syedsalman7028
    @syedsalman7028 3 месяца назад

    I was really struggling in understanding fine tuning and can't find proper tutorials for it but from this video I am positive that after the end of this playlist I will understand fine-tuning properly.

  • @BalaguruGupta
    @BalaguruGupta 6 месяцев назад

    Krish, you are making us also EVOLVE along with you :-) Thanks a lot for this beautiful explanation.

  • @sivakumar-ho3mw
    @sivakumar-ho3mw 8 месяцев назад

    Really awesome content and as always no words to say about your teaching skill.Thanks for sharing the knowledge

  • @abduljalilhassan1798
    @abduljalilhassan1798 9 месяцев назад +2

    We are waiting for part 2

  • @avanthikar2608
    @avanthikar2608 9 месяцев назад +2

    Can you please upload videos indepth of how different prompting techniques like chain of thought, self consistency, knowledge generation etc were practically used with which the outputs of the models based on use cases are getting improved..
    I am requesting for so long time please make videos on these topics please

  • @Rohit_Rai901
    @Rohit_Rai901 9 месяцев назад +3

    Sir! make whole playlist on Data Science end to end
    Sir make world best Data Science course on youtube cover all topics
    We are waiting🥱🥱

  • @pratibhagoudar6817
    @pratibhagoudar6817 9 месяцев назад +2

    Sir, please make a video on the model pruning technique for llms

  • @velugucharan8096
    @velugucharan8096 9 месяцев назад +1

    Please upload the fine-tune video fast as much as possible sor

  • @Nileshkumar-lf7oc
    @Nileshkumar-lf7oc 9 месяцев назад +3

    Hey krish you mentioned that you learnt generative ai theories 2-3 months ago. Could you please share the resources you used during your learning journey? Any guidance would be greatly appreciated!

  • @mubasharsaeed6044
    @mubasharsaeed6044 9 месяцев назад +2

    Please make a video llm for robotics please with practical

  • @TherajLal
    @TherajLal 7 месяцев назад

    very nice session, learned to much about quantization.

  • @ay-tw4om
    @ay-tw4om 9 месяцев назад +2

    hi can you tell whether training custom embeddings improves performance?

  • @kashifsadiq4145
    @kashifsadiq4145 9 месяцев назад +2

    Great

  • @eagleblaze9292
    @eagleblaze9292 9 месяцев назад

    Waiting for part 2 sir❤

  • @ADARSHTenZo
    @ADARSHTenZo 9 месяцев назад +1

    Thank you sir

  • @saidhanush703
    @saidhanush703 9 месяцев назад +2

    Krish iam having a doubt , if i have fintunned a model with my custom data , and my data will be keep on adding for every week , so in that case do i need to finetune again and again every week ?

  • @mohsenghafari7652
    @mohsenghafari7652 9 месяцев назад +2

    hi. please help me. how to create custom model from many pdfs in Persian language? tank you.

  • @rambabugudepu1263
    @rambabugudepu1263 9 месяцев назад +2

    Hi sir good morning I have one doubt you will help me. I am practicing cotton disease project. I will give predict without cotton images I want show "there is no image"

  • @ChinmayRaut-no5zu
    @ChinmayRaut-no5zu 7 месяцев назад

    Sir, This was a better explaination of Quantization. But why this is not covered in your paid course of Mastering LLM on Ineuron? In that course, only a high level overview is given without the mathematical intution.

  • @aswinrajsharma2244
    @aswinrajsharma2244 9 месяцев назад +3

    ❤❤❤ thank

  • @himanshurathod4086
    @himanshurathod4086 9 месяцев назад +2

    krish can you please make video on kubernetes like docker?

  • @IdPreferNot1
    @IdPreferNot1 8 месяцев назад

    Excellent series

  • @rohanpandey9957
    @rohanpandey9957 9 месяцев назад +1

    sir i actually had a question what are the pre-requisites to understand each and every bit of fine tuning these large language models? can you point me to some of your resources?

  • @ku_lifestyle2639
    @ku_lifestyle2639 9 месяцев назад +2

    What is the software name in your M2 sir. for the writing you are using?

  • @TravelWithFunn
    @TravelWithFunn 9 месяцев назад +2

    Hello sir, I am searching for an internship related to data science, data analytics, machine learning...
    If there is any platform form or organization then please let me know.
    It will be very helpful for me.

  • @rakeshkumarsharma195
    @rakeshkumarsharma195 2 месяца назад

    Awesome 💯

  • @panditsingh1302
    @panditsingh1302 9 месяцев назад +1

    Hello Krish sir good Evening
    Why ineuron support team is not responding ?
    I'm Trying from last Friday regarding neuro lab and from last three days regarding assessment evaluation but no one responding.

  • @codewithtej494
    @codewithtej494 9 месяцев назад

    PTQ and QAT is not clear.
    PTQ: Train, then shrink (fine-tune first, then quantize).
    QAT: Train with a smaller (combine fine-tuning and quantization).

    • @priyankjain9970
      @priyankjain9970 6 месяцев назад

      u r right. In both cases Quantization is happening which will lead to loss of info. In QAT he mentioned that there is no loss of info. Couldn't understand how

  • @vishalaiml1649
    @vishalaiml1649 7 месяцев назад

    Hi Krishnaik, can you please create a Series on securing LLM responses, and Guardrails as it is burning topic now a days. Sincere Request.

  • @domvito55
    @domvito55 10 дней назад

    Hi, sir...There is a mistake when you explain what exponent and mantissa are.
    exponent is not the integer value. that is how many times you will multiple by 10. In this case, the number would be stored as 10 * 0.732

  • @SameraSamera1444
    @SameraSamera1444 7 месяцев назад

    Amazing many thanks 🙏

  • @ManujJoshi
    @ManujJoshi 8 месяцев назад

    good explanation

  • @sumayaabdulrahman471
    @sumayaabdulrahman471 Месяц назад

    Question : can i watch this playlist with basic knowledge of langcahin and RAG or do i need to watch krish naiks 'updated langchain' playlist to understand this playlist?

  • @demoacct0121
    @demoacct0121 8 месяцев назад

    May I know what app you are using to perform the writing and drawing on the screen?

  • @nethmidharani2281
    @nethmidharani2281 9 месяцев назад

    Do the remaining vedios about qlora lora peft asap please please please🙏

  • @KumR
    @KumR 5 месяцев назад

    Can u please add that first video in this same play list?

  • @umarfarooque3687
    @umarfarooque3687 6 месяцев назад

    How do we handle hallucination when finetuing is done ?

  • @ammy490
    @ammy490 8 месяцев назад

    Is QAT not applicable on pre trained model??

  • @intrepidani6422
    @intrepidani6422 6 месяцев назад

    hey everyone,
    I'm Ani,
    and im seriously confused few things and i need a help from anyone please,
    So i want to create a good AI models so I have started with learning AI/Ml I have some codeing experience before so iwas learning DSA but from 1.5 months i'm sutck in it. today when i was watching this video i got my interest back but the thing is i don't have any Practical experience in it so i dont really understand it properly, can you please clarify what must i do shall i continue learning this or i must first do Mechine Learning then come back to this

  • @Dream-fm6dq
    @Dream-fm6dq 7 месяцев назад

    anyone can tell me quickly how can i get only response from llama 2 response. it give first input and then response but i want only response

  • @saifulislamsarfaraz
    @saifulislamsarfaraz 20 дней назад

    What is the software name of your notepade?

  • @mehul4mak
    @mehul4mak 3 месяца назад

    pdf of such notes are highly appreciated.

  • @HarshPatel-sw1jq
    @HarshPatel-sw1jq 7 месяцев назад

    I feel like exponent ad mentissa is wrong in the explaination

  • @Radhe_Krishna2245
    @Radhe_Krishna2245 9 месяцев назад +2

    Kyu pda rhe ho giru g iska kya fyada h.tabhi to hm dekhenge

  • @studology67
    @studology67 5 месяцев назад

    Sooo many adds in your video,

  • @aiforeveryone
    @aiforeveryone 9 месяцев назад +1

    Great