No video :(

Chat With Documents Using ChainLit, LangChain, Ollama & Mistral 🧠

Подписаться 13 тыс.

Просмотров 12 тыс.

50% 1

In this video, I am demonstrating how you can create a simple Retrieval Augmented Generation UI locally in your computer. You can follow along with me by cloning the repo locally. You can also use LangSmith for tracing the LLM calls and use LangChain Hub for using already available prompt template for different models, for this case, mistral.
Open Source in Action 🚀
- Mistral is used as Large Language model.
- LangChain is used as a Framework for LLM
- Mistral model is downloaded locally using Ollama
- Chainlit is used for deploying.
👉🏼 Links:
Chainlit: docs.chainlit....
LangChain: www.langchain....
Ollama: ollama.ai/
💻 GitHub repo for code: github.com/sud...
------------------------------------------------------------------------------------------
☕ Buy me a Coffee: ko-fi.com/data...
✌️Patreon: / datasciencebasics
------------------------------------------------------------------------------------------
🔗 🎥 Other videos you might find helpful:
🔥 Databricks playlist: • 30 Days Of DataBricks
⛓️ Langflow: • ⛓️ langflow | UI For 🦜...
⛓️ Flowise: • Flowise | UI For 🦜️🔗 L...
🔥Chainlit playlist: • Chainlit
🦜️🔗 LangChain playlist: • LangChain
🦙 LlamaIndex Playlist: • LlamaIndex
------------------------------------------------------------------------------------------
🤝 Connect with me:
📺 RU-vid: www.youtube.co...
👔 LinkedIn: / sudarshan-koirala
🐦 Twitter: / mesudarshan
🔉Medium: / sudarshan-koirala
💼 Consulting: topmate.io/sud...
#langchain #chainlit #ollama #mistral #rag #retrievalaugmentedgeneration #chatgpt #datasciencebasics

Опубликовано:

28 авг 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 42

@bp0819 6 месяцев назад

Love it thanks! Been fighting a bunch of programs to get open models to work and this saved my life!

@datasciencebasics 6 месяцев назад

You are welcome. Flad that you find it helpful !!

@mohamedemarah9379 5 месяцев назад

Thank you so much ! Watched a lot of videos online and yours was the only that helped me get it up and running :)

@datasciencebasics 5 месяцев назад

You are welcome, Glad that it was helpful :)

@genix4897 3 месяца назад

i am not able to run it, when i upload a document, it just keeps processing forever please help me

@NehaH-u6y Месяц назад

did you solve it

@chibuzoemelike6403 6 месяцев назад

Thank you so much. Seeing the possibilities of LangChain and ollama is so great. Is it possible to connect this with a database, say bigquery so that you could chat with the database or table and get responses. It would be great to see a video on that. Thank you so much for your great work.

@datasciencebasics 6 месяцев назад

You are welcome. If I find something what you mentioned, will share it via video !!

@thedefende-i7g 6 месяцев назад

i have been asked to build poc on document search: i have used streamlit and build the simple ui application using google ai api and gemini pro.. now i have to implement below functionality: "Should have design & approach of the solution Two types of user to use this document search bot. Admin & Normal User. Admin should be able to upload and delete documents. Both user should be able to ask query The response should match the query asked Have test cases created for validation" please help me out with the source code how to start and build

@mrweeed5066 6 месяцев назад

Thank you, subbed to your channel. Can you also please give a tutorial on other topics of chainlit like Chat Generation, Elements (Image, Pyplot, TaskList etc.,) and AskUser section. I'd love to see more of your videos on "ChainLit, LangChain, Ollama & Mistral" series, they are easy and very helpful. and loading previous complete chat interaction on chainlit will be interesting video. I'm also trying to connect my React frontend app to chaint using this combination, I'm stuck at changing chat profiles from React frontend to be able to use Mistral and GPT4all at when selected from tab, if you can make a video on this, its also very helpful and again thank you for the video... :)

@datasciencebasics 6 месяцев назад

You are welcome. Will try my best to create more more similar videos !!

@achrafrahime3347 3 месяца назад

what about this error : Collection langchain is not created.

@genix4897 3 месяца назад

were you able to solve it?

@IamalwaysOK 6 месяцев назад

Thank you for you videos. They are very informative. Could you please also create a video that explains how to use "BioMistral" with Ollama?

@datasciencebasics 6 месяцев назад

You are welcome. You can create your own model using Ollama by downloading model from Huggingface. You can take help of this videa as reference Ollama: How To Create Custom Models From HuggingFace ( GGUF ) ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-TFwYvHZV6j0.html

@IamalwaysOK 6 месяцев назад

@@datasciencebasics Thank you for your response. Indeed, I watched that video and found it highly informative. However, it appears that the Biomistreal quantized model operates differently compared to the quantized model by The Bloke. According to Ollama, the creation of the model from the model file was successful, but attempting to converse with it proved unsuccessful. It consistently returns only the start token.

@datasciencebasics 6 месяцев назад

hmm.. might create a video on it. Let’s see .

@IamalwaysOK 6 месяцев назад

@@datasciencebasics Thank you very much for the information. Just so you know, in the medical field, Ollama offers a model called Meditron. I'm interested in hearing your thoughts on how Meditron compares to bioMistral. Additionally, for someone who is new to this area, how would you recommend they evaluate these models? Given the rapid release of new open-source models, is there a universally accepted method for comparing them?

@datasciencebasics 6 месяцев назад

Which one to choose depends upon ones usecase and also trail and error is needed to choose the best one. Domain knowledge is key in using such models. Here is the link for running biomistral locally using Ollama. How To Create Medical Chatbot Using BioMistral, Ollama, LangChain & Chainlit ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-bANziaFj_sA.html

@dprasen 5 месяцев назад

You saved my life, can you provide 1-1 sessions?

@datasciencebasics 5 месяцев назад

Glad it was helpful. For coaching / consulting, here is the link. topmate.io/sudarshan_koirala

@MarcvitZubieta 6 месяцев назад

🎯 Key Takeaways for quick navigation: 00:00 🛠️ *Fourth video in an AMA series demonstrating creation of a simple Retrieval Augmented Generation UI using various models.* 01:07 🧩 *Utilizing AMA, Lang chain, ChainLit, and Mistral for deploying applications and tracing.* 03:25 📦 *Setting up a virtual environment and installing necessary packages for the project.* 06:43 📜 *Installing AMA models locally for use in embedding and generating responses.* 09:54 🤖 *Implementing a chat interface for interacting with documents, leveraging retrieval QA chains.* 10:48 🖥️ *Running and interacting with the created application to ask questions related to uploaded PDF files.* 14:04 🔄 *Exploring additional functionalities for uploading PDF files directly onto the UI for interaction.* 19:36 🎥 *Encouragement to like, share, and subscribe for more content, with a focus on practical application and experimentation.* Made with HARPA AI

@datasciencebasics 6 месяцев назад

So this AI just gets AMA for Ollama 😜

@atrocitus777 6 месяцев назад

How can you enable streaming and render the sources appropriately. put an issue on your github at my attempt at this.

@fighter4215 6 месяцев назад

When running the ingest.py I am getting this error : ValueError: Error raised by inference API HTTP code: 404, {"error":"model 'mistral' not found, try pulling it first"}. Help please

@datasciencebasics 6 месяцев назад

hello, did you pull the mistral model using Ollama ? The error suggests you to download model first.

@fighter4215 6 месяцев назад

@@datasciencebasics thank you it was that !

@drmetroyt 6 месяцев назад

Please include installation on docker

@kaleshashaik5959 5 месяцев назад

Thank you. how to run this in offline mode locally ? I mean, without internet.

@datasciencebasics 5 месяцев назад

You are welcome. After downloading model via Ollama, you can run it locally offline.

@kaleshashaik5959 5 месяцев назад

@@datasciencebasics Thanks. I am creating one new application using Django and from admin UI, data will be inserted to postgres database. Now, I want to use any private/local LLM (without internet) to create a chat application that query the database by user question. Is this can be achieved ? Can you share any reference on this ?

@AI_GB 6 месяцев назад

what customization are you using in your terminal?

@datasciencebasics 6 месяцев назад

this palylist might be helpful Productivity ru-vid.com/group/PLz-qytj7eIWWFKENNKJF_7nvmYunOFn0u

@Alkotas25 6 месяцев назад

hi, is this solution is private? so all data, vectors, llm, etc. run locally?

@datasciencebasics 6 месяцев назад

hello, yes all running locally.

@Alkotas25 6 месяцев назад

@@datasciencebasics good, thx, I try it.👍

@dgm1949 6 месяцев назад

please see my comment regarding being online.

@shashankpandey8404 5 месяцев назад

from where do i get the langchain api ?

@datasciencebasics 5 месяцев назад

refer to this video LangSmith For Beginners | Must know LLM Evaluation Platform 🔥 ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-FgG-trkAMwU.html

@sunilanthony17 3 месяца назад

chain.acall(message.content, callbacks=[cb]) chain.acall is now deprecated and ainvoke does not work with chainlit because it doesn't you that it's process "ReterivalQA and show the steps.

@NehaH-u6y Месяц назад

the file keeps processing forever. How to solve this? please help