Mini project on RAG using Open source LLM hosted on GROQ cloud platform - Chatbot with PDF files - PDF Conversation with RAG_Another LLM.py
This project is to establish a conversational chat with the PDF files. The PDF files will be providing the context for the user query.
This is developed using the langchain framework, open source LLM - llama3-groq-70b-8192-tool-use-preview (this can be changed based on the evaluation).
The solution involves the below steps,
- Data ingestion - PDF files.
- Splitting the PDF files into chunks and convert them into the embeddings using open source embeddings offered by Huggingface.
- Store the embeddings into the vector database - Chroma.
- Define chat template as offered by Groq.
- Select the Open source model offered by Groq - https://groq.com/
- Define the chain with retriever.
- Take the input from the user - question/query from the end user.
- Generate response and maintain the chat history using the ChatMessage History provided by Langchain.
- Streamlit is used to provide a simple web application.
- Langsmith is used to monitor the application
Special thanks to Krish Naik for providing a detailed session on developing this.