Available for work

LLM Based Q/A on Web Context - LLM Q/A based on OLLAMA

Date

December 2024

Service

Employer

Project Overview

This project aimed to build an interface so that user could provide an URL and ask specific questions to the LLM based on that webpage where the URL leads. The LLM also indicated the paragraphs and texts that led to the answer that it provided. This reduced the chances of hallucinations by the model. The user could also see the answer the model gave based on its own knowledge and could choose the one that suited him the most.

https://github.com/ritam3/LLM-Based-Q-A-based-on-Web-Context

Key Highlights

  • Used Langchain to work with LLMs and loading web based data, splitting text and converting them into documents.

  • FAISS to store the encodings.

  • Worked with OLLAMA to include models like LLAMA 3.2 and Gemma:2b.

  • Used Streamlit to build the User Interface.

  • User was given the ability to ask a question and the chosen LLM model returned the answer based on web context and its own knowledge.

More Project

RAG-PDF QA Method

Implementation of RAG method to perform question answering on PDF documents. HuggingFace embeddings in addition to the Groq API were used for efficient and accurate information retrieval. Session management was done to ensure embeddings are computed once per document upload to optimize performance by avoiding redundant computations.

January 2025

RAG-PDF QA Method

Implementation of RAG method to perform question answering on PDF documents. HuggingFace embeddings in addition to the Groq API were used for efficient and accurate information retrieval. Session management was done to ensure embeddings are computed once per document upload to optimize performance by avoiding redundant computations.

January 2025

RAG-PDF QA Method

Implementation of RAG method to perform question answering on PDF documents. HuggingFace embeddings in addition to the Groq API were used for efficient and accurate information retrieval. Session management was done to ensure embeddings are computed once per document upload to optimize performance by avoiding redundant computations.

January 2025

Next Word

"Forecasting the Future: Utilizing LSTM for Next-Word Prediction in Natural Language Processing."

December 2024

Next Word

"Forecasting the Future: Utilizing LSTM for Next-Word Prediction in Natural Language Processing."

December 2024

Next Word

"Forecasting the Future: Utilizing LSTM for Next-Word Prediction in Natural Language Processing."

December 2024

Churn Prediction and Customer Cohort

Development of a Machine Learning Model that takes user's activity data as input and predicts if the user is going to churn

October 2023

Churn Prediction and Customer Cohort

Development of a Machine Learning Model that takes user's activity data as input and predicts if the user is going to churn

October 2023

Churn Prediction and Customer Cohort

Development of a Machine Learning Model that takes user's activity data as input and predicts if the user is going to churn

October 2023