Introduction:

This project implements a LLM based Chatbot using Transformers and Langchain. It is specifically used for assisting the users planning to apply to Karunya.

Key Features

• Developed a chat interface using Streamlit.
• Scraped the relevant data from Karunya's website and stored it in a MySQL database.
• Included support for vernacular languages.
• Utilised a RAG system for data withdrawal from a vector database using langchain
• Added functionality for the chatbot to remember previous chats.
• Implemented a NER (Named Entity Recognition) based user data collection for Vernacular Language Chat.
• Implemented a personalized style based on the field of interest of the user.
• Implemented TTS with vernacular support

Getting Started

1. Clone the repository:

git clone https://github.com/kevin-291/llm-chatbot.git

2. Install dependencies:

pip install -r requirements.txt

3. Running Scripts:

streamlit run app.py

• Do ensure that streamlit is installed in your system before running the script.

4. Further Development:

• Incorporating vLLM for a high-throughput and memory-efficient inference and serving engine for LLMs.
• Incorporating STT with support for vernacular languages.