L-9 Build a Q&A App with RAG, LangChain, and Open-Source LLMs | Step-by-Step Guide

Поділитися
Вставка
  • Опубліковано 31 січ 2025

КОМЕНТАРІ • 52

  • @sangeethag8228
    @sangeethag8228 2 місяці тому

    Thanks a lot, Madam !!!. You are awesome at Explaining things in a very calm and simple way, whereas some UA-camrs exaggerate. :)

  • @omkarsatapathy8209
    @omkarsatapathy8209 6 місяців тому +1

    Thank you so much, madam. You really considered my comment regarding RAG implementation without any secret key. Thank you so much again.. keep posting and keep growing !! I will definitely share this video with all my Network. Happy coding !

  • @xiaochen-l7c
    @xiaochen-l7c Місяць тому

    I hope there will be more basic concept courses and practical courses. I like the way you tell you very much, and your teaching is very vivid

    • @CodeWithAarohi
      @CodeWithAarohi  Місяць тому

      Thanks for the feedback! I'm planning on making more content like this.

  • @itsmeuttu
    @itsmeuttu 6 місяців тому

    Very Helpful
    keep it up mam !!

  • @arnavthakur5409
    @arnavthakur5409 4 місяці тому

    Very helpful for me

  • @xiaochen-l7c
    @xiaochen-l7c Місяць тому

    very nice

  • @soravsingla8782
    @soravsingla8782 4 місяці тому

    Good work

  • @neelamadhabkhaya98
    @neelamadhabkhaya98 6 місяців тому

    I have only 1 like option. Again and again try to like these videos. Really helpful.

  • @GradientPlayz
    @GradientPlayz 5 місяців тому

    You are the best

  • @sangeethag8228
    @sangeethag8228 2 місяці тому

    If I have the option to subscribe 1M times, I will do so. But for one ID, there is only one subscription. U r awesome !!!!

  • @KPBhan
    @KPBhan 6 місяців тому

    Very nice and thank you very much.
    please help with training models ways or examples

  • @alexramos587
    @alexramos587 3 місяці тому

    Thanks.

  • @ma8695t
    @ma8695t 7 днів тому

    I wish you could let us know how to get a clean or refined final answer

  • @devavratpro7061
    @devavratpro7061 3 місяці тому

    Thanks Aarohi Mam for your valuable video. Can I change prompt response format in such a way that it fills the details in a fixed template:, like, filling tender fields from the requirements/specifications in a pdf file? Please guide further through the details!

  • @Samaya_malik88
    @Samaya_malik88 Місяць тому +1

    Mam can you make video on api callin or tool calling or function calling how to do .using local LLM longchain and rag dataset

  • @hendoitechnologies
    @hendoitechnologies 6 місяців тому

    post regular video about Generative AI - full course

  • @SumaiyaJahan-y4c
    @SumaiyaJahan-y4c 6 днів тому

    I wanted to know about which configuration you are using?Because I am facing out of memory issue for this task. I have 1TB CPU, 2GB GPU.If I use CPU it takes more than 1hr to complete for just 1 question,on the other hand if I use GPU it is causing memory issue!!

  • @zeadmostafa5100
    @zeadmostafa5100 6 місяців тому

    Thank you very much for your effort.
    I want to ask you if i can use colab or kaggle notebook instead of running the code in my local machine ?

  • @thomaseldhose1006
    @thomaseldhose1006 5 місяців тому

    How can I convert my unstructured data into structured data?

  • @hendoitechnologies
    @hendoitechnologies 6 місяців тому

    post video about how to fine-tune "Claude 3.5 sonnet API" - full course video for developers..please

  • @veerabhadrayyakalacharanti4051
    @veerabhadrayyakalacharanti4051 2 місяці тому

    hello mam i am stuck with output called Loading checkpoint shards its not loading only, its strck like that only, is it take much time to lode or any problem or what i have to do

  • @NidhiDixit-fr4ew
    @NidhiDixit-fr4ew Місяць тому

    How do we know which huggingface embedding I have to download? please somebody help me

    • @CodeWithAarohi
      @CodeWithAarohi  Місяць тому

      It depends upon the type of task you are performing and domain of the data.
      If you are performing general text processing or you are unsure about the specific domain. Use general-purpose models like all-MiniLM-L6-v2 or all-mpnet-base-v2. These are fast, lightweight, and work well for a broad range of NLP tasks.
      If your data is specific to a certain domain (e.g., legal, scientific, financial, etc.), select a model fine-tuned for that domain.
      Legal texts: Use models like nlpaueb/legal-bert-base-uncased.
      Scientific texts: Use models like allenai/scibert_scivocab_uncased.
      Financial texts: Use sentence-transformers/finbert.

    • @NidhiDixit-fr4ew
      @NidhiDixit-fr4ew Місяць тому

      @@CodeWithAarohi Thank you so much for your time.. please start taking classes where we can ask you our doubts..

  • @salah5894
    @salah5894 5 місяців тому

    please it said that i have to download the model and it took 9.5 GB , is it true ? or there is other method without downloading it

    • @CodeWithAarohi
      @CodeWithAarohi  5 місяців тому

      You need to download the pretrained model. You can try using some other LLM which is smaller as compare to this model.

    • @salah5894
      @salah5894 5 місяців тому

      @@CodeWithAarohi OK thank u so so so much 💗💗

  • @arshmaanali714
    @arshmaanali714 4 місяці тому

    Mam please create an intelligent chatbot using Streamlit and Langchain (RAG), where the
    chatbot can receive voice input in Urdu/Hindi, process it, and return both text and audio responses in
    Urdu/hindi. The chatbot should be able to interact with users fluently, allowing for seamless
    audio-to-text and text-to-audio communication in the Urdu/Hindi
    Workflow:
    ● Build the Streamlit interface for real-time Urdu/hindi audio input and output.
    ● Integrate Langchain (RAG) with an LLM (Language Model) API to generate dynamic responses
    based on the user’s input. (use PDF files only)
    ● Ensure the chatbot responds not only with a text-based answer in Urdu/Hindi but also converts that
    response back to audio and plays it for the user language.

  • @amalkuttu8274
    @amalkuttu8274 6 місяців тому

    iam not able to laod the huggingfaceembeddings. it shows this error.
    The specified module could not be found. Error loading "C:\Users\aj441\anaconda3\envs\llmenv\lib\site-packages\torch\lib\fbgemm.dll" or one of its dependencies.

    • @CodeWithAarohi
      @CodeWithAarohi  6 місяців тому

      @@amalkuttu8274 are you running through anaconda or cmd prompt

    • @amalkuttu8274
      @amalkuttu8274 6 місяців тому

      @@CodeWithAarohi anaconda

    • @eashan2405
      @eashan2405 5 місяців тому +1

      I also faced the same issue. Intsead of using the requirements.txt file, just install them directly using these commands:
      1. conda create -n env_langchain2 python=3.10
      2. conda activate env_langchain2
      3. conda install pytorch torchvision torchaudio cpuonly -c pytorch
      4. pip install transformers
      5. pip install sentence-transformers
      6. pip install langchain langchain_community langchain-huggingface langchain_experimental langchain_chroma langchainhub
      7. pip install streamlit
      8. conda install jupyter
      9. jupyter notebook
      Then test your installation by running this script in Jupyter Notebook:
      import torch
      import transformers
      import sentence_transformers
      import langchain
      print("PyTorch version:", torch.__version__)
      print("Transformers version:", transformers.__version__)
      print("Sentence Transformers version:", sentence_transformers.__version__)
      print("LangChain version:", langchain.__version__)
      It worked for me! Let me know if you still face issues.

    • @amalkuttu8274
      @amalkuttu8274 5 місяців тому

      @@eashan2405 I will surely look that.

    • @Rohitpampatwar
      @Rohitpampatwar Місяць тому

      @@eashan2405 Really good one