readme.md

Samsung Manual Chatbot

A chatbot built with LangChain, Gradio, and Hugging Face Transformers that allows you to interact with the Samsung Manual.
It uses embeddings + ChromaDB for retrieval and a conversational chain for contextual Q&A.

---

Project Structure

LLM_chatbot2/

│── chroma_db/ # Persistent Chroma vector database

│── temp_docs/ # Store documents (Samsung manual here)

│── app.py # Main Gradio app

│── requirements.txt # Dependencies

│── README.md # Project documentation

│── .gitattributes

---

Features

• Document Loading – Loads and processes the Samsung manual (temp_docs/samsung_manual.txt).
• Chunking – Splits the document into manageable chunks for embeddings.
• Embeddings – Uses sentence-transformers/all-MiniLM-L6-v2 for semantic search.
• Vector Database – Stores embeddings in ChromaDB for retrieval.
• Conversational Memory – Maintains chat context with ConversationBufferMemory.
• LLM Response – Powered by google/flan-t5-base via Hugging Face pipeline.
• Gradio UI – Simple chat interface for interacting with the chatbot.

---

Installation

Clone the repository and install dependencies:

git clone https://github.com/Anvit25/LLM_chatbot2.git
cd LLM_chatbot2
pip install -r requirements.txt
Running the App
Make sure you have the Samsung manual text file at temp_docs/samsung_manual.txt.

Then run:

python app.py
Gradio will launch a local server. Open the link shown in the terminal (usually http://127.0.0.1:7860) to interact with the chatbot.

Requirements

Dependencies are listed in requirements.txt:

pypdf gradio langchain chromadb sentence-transformers transformers torch

Usage

Place your document inside temp_docs/ (default: samsung_manual.txt).

Run the app.

Ask natural language questions like:

"How do I reset my Samsung washing machine?"

"Explain the safety precautions in the manual."

The chatbot retrieves relevant chunks from the document and answers conversationally.

Customization

Change the document: Replace temp_docs/samsung_manual.txt with any .txt file.

Switch embeddings: Modify MODEL_NAME_EMBEDDINGS in app.py.

Try different LLMs: Update MODEL_ID_LLM to another Hugging Face model.

License

This project is open-source under the MIT License.

---