Mistral-7B-Instruct Multiple-PDF Chatbot with Langchain & Streamlit |FREE COLAB|All OPEN SOURCE #ai

Join me in this tutorial as we explore the development of an advanced Chatbot for handling multiple PDF documents, harnessing the power of open-source technologies. We'll dive into the world of Sentence Transformers for text embeddings, make use of Faiss CPU for efficient vector storage, and seamlessly integrate the Mistral-7b model. In addition, we'll employ the Streamlit library to create a dynamic conversational interface, enabling you to embark on a step-by-step journey to build a smart and effective PDF Document Chatbot, all while leveraging readily available resources.

To make this even more accessible, we'll utilize the free version of Google Colab to create a demo of our chatbot. Afterward, we'll transition to building a Streamlit application for a practical and user-friendly interface. Whether you're new to this field or an experienced enthusiast, this tutorial will enable you to delve into the realm of conversational AI and drive innovation.
#ai #llm #largelanguagemodels #generativeai #streamlit #mistral #opensource

[DataEdge]

(turn on notifcation bell), like, and comment what else you want to see! Love y'all and have an amazing day.Thank you so much!

[jaytrivedi]

Is mixtral faster then llama-7b-gguf??
Specially on cpu…. I do Have 32gig of RAM but no gpu…which gguf model should I use to get faster response.??

[matmorales]

what are you specs of your computer? and how long it takes per question? I am asking because one file has 50 pages and another file has 17 pages.

[darshanjoshi]

Hi! When I try to upload new PDF, the chat history remains as it is. What is I wanted to start a whole new conversation interface when new PDF is uploaded?

[ashvathnarayananns]

How to add Prompt to our local model, when adding a prompt i am facing token limit exceeded 512. can you post a video of "PDF reader with prompt"

[smarhl]

Nice video, Im trying to do something similar to yours but my inference times are in the range of 2-3 minutes on a 8th gen i7 cpu. Have you experienced this while testing?

[kritika-wk]

As you have a single path to mention in model_path, i have many, like in model folder, i have 3 shards of the fine tuned model, along with the tokenizer files. Its giving me error when i am providing it in the list, please help fast as you can, as i am stuck for long.

[BestRestaurantToEat]

Hi, I am using VS Code instead of Colab, is there any difference to the program and what we need to change. TQ

[akshay]

Very nice and informational video for usage. I just had several questions and would be grateful to you if you could answer them.
1. Where can and should we save our vector embeddings and also if I want it to be free what should I use.
2. I want to remove the uploading pdf part and directly use a Q and A interface with model fine tuned with my PDF's and Text documents, what steps should I follow or remove in this?
3. I want to deploy this chatbot in a server and make a full fledged website go online to display in my Projects. What methods can I use to implement it.
Please Help me in these Questions. 🙏🙏🙏

[prathampendam]

Hi, I am facing this error : Number of tokens (2121) exceeded maximum context length (512). Can you help me with this

[taison]

Can we deploy the complete application as a web app on azure and get good performance? Can you guide how

[pjlfomn]

Hyy @DataInsightEdge, your video really helped me a lot and seeing your replies to questions, I also have some questions hoping to get answer as soon as possible:
1. can streamlit work if a upload this code to a server as it is or i need to built the ui with something else?
2. what to do in case i want to train mistral with my personal repository (almost containing java files) instead of pdfs?
3. this mistral instruct 7b model requires how much computation at what size of data?
4. Which will be more faster to use and how much to be used in case of large repository, gpu, tpu, cpu..

[ShreyaSaxena-koli]

will this answer questions related to tables present in pdf ?

[seinaimut]

thanks for your tutorial. can i use txt file not only pdf?

[mahshid-Engineer]

Thank you for shared the video. I very reflect the find one away the little limit for summery PDF white the one page after the read page one the summery and send next agent. the all the try the work to arrive end page of pdf. i dont no the doing is the work. Are you help me?

[vragel]

Thanks for explanation. I want to use my english pdf document with your chatbot app. But can i ask turkish questions and take turkish answers on my english pdf with your model?

[matthieugillieron]

how do you change avatar_style with a image.png?
thanks

[Cosmo.Curious]

sir is it possbile to use a smaler model like phi 1.5 or stable lm 3b

[GECBE]

Thanks for this amazing tutorial. Can you please tell me how can I host RAG as complete web app on Google cloud?
Please

[chidanandkumarks]

Can you share sample data you have worked on. This will help us in replication your work given in colab notebook. Do the needful