|
--- |
|
title: ChatBot for Jupyter Notebook |
|
emoji: 💬 |
|
colorFrom: green |
|
colorTo: purple |
|
sdk: gradio |
|
app_file: app.py |
|
pinned: false |
|
license: apache-2.0 |
|
short_description: ChatBot using Microsoft's Phi-3-mini-4k-instruct |
|
--- |
|
|
|
# ChatBot for Jupyter Notebook |
|
|
|
## Introduction |
|
Welcome to the **Chatbot for Jupyter Notebooks project!** This repository contains an interactive chatbot built using **Microsoft's Phi model(Phi-3-mini-4k-instruct)**, integrated with Jupyter Notebook. The chatbot is designed to assist users by providing relevant answers and insights based on the code and content present in the notebook. |
|
|
|
The primary goal of this project is to enhance the user experience by enabling an intuitive and efficient way to interact with the notebook's code. |
|
|
|
### Key Features: |
|
- **RAG Framework:** Retrieval Augmented Generation Framework has been implemented to get the relevent context related to queries that deals with referring a particular cell number, or with queries asking follow-up questions with respect to previous responses from the model. |
|
- **Context-Aware Responses:** By analyzing the code and content in the notebook, the chatbot provides contextually relevant answers. |
|
|
|
## Usage |
|
Upload your Jupyter Notebook and start asking questions. Examples for the questions that you can ask for: |
|
- Explain the code present in 4th cell. |
|
- Can you summarize the above/previous response |
|
|
|
### Few things to remember: |
|
- Before uploading your notebook, make sure that the output of the cells are cleared, as it can cause errors occasionally. |
|
- If the notebook has MarkDown cells, it can sometime disturb the order of the cells that the model gets as context for answering the query, |
|
so it's better to avoid MarkDown cells if any. |
|
- The context has been provided in a way to answer the queries by just referring the cell number, But sometimes it might not be accurate to pick up the right cell number. |
|
|
|
|
|
|
