Spaces:
Sleeping
Sleeping
update readme file
Browse files
README.md
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
---
|
| 2 |
title: Ragent Chatbot
|
| 3 |
-
emoji:
|
| 4 |
colorFrom: blue
|
| 5 |
-
colorTo:
|
| 6 |
sdk: gradio
|
| 7 |
sdk_version: 5.36.2
|
| 8 |
app_file: app.py
|
|
@@ -11,4 +11,55 @@ license: mit
|
|
| 11 |
short_description: A Smart AI chatbot powered by RAG and AGENT
|
| 12 |
---
|
| 13 |
|
| 14 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
title: Ragent Chatbot
|
| 3 |
+
emoji: π€
|
| 4 |
colorFrom: blue
|
| 5 |
+
colorTo: green
|
| 6 |
sdk: gradio
|
| 7 |
sdk_version: 5.36.2
|
| 8 |
app_file: app.py
|
|
|
|
| 11 |
short_description: A Smart AI chatbot powered by RAG and AGENT
|
| 12 |
---
|
| 13 |
|
| 14 |
+
## π¬ Ragent Chatbot Preview
|
| 15 |
+
|
| 16 |
+
 <!-- Add your app preview image in 'figure/thumbnail.png' -->
|
| 17 |
+
|
| 18 |
+
# π€ Ragent Chatbot
|
| 19 |
+
|
| 20 |
+
**Ragent Chatbot** is an intelligent **retrieval-augmented agent assistant** powered by LLMs. It combines the power of **RAG (Retrieval-Augmented Generation)** with **agent-based tool reasoning**, allowing it to dynamically respond using retrieved knowledge or external tools depending on the query.
|
| 21 |
+
|
| 22 |
+
## π§ What It Can Do
|
| 23 |
+
|
| 24 |
+
- Answer user questions by retrieving information from a custom document store
|
| 25 |
+
- You can upload any document and then ask the chatbot to retrieve answer of your question.
|
| 26 |
+
- Automatically decide when to use tools (like search, calculator, etc.)
|
| 27 |
+
- Combine multiple steps of reasoning using the ReAct agent pattern
|
| 28 |
+
|
| 29 |
+
## π Features
|
| 30 |
+
|
| 31 |
+
- π **Hybrid Search**: Combines vector similarity and BM25 keyword matching for relevant document retrieval
|
| 32 |
+
- π€ **ReAct Agent**: Uses tool-based reasoning when knowledge retrieval is insufficient
|
| 33 |
+
- π¬ **Gradio Chat UI**: Simple and responsive chat interface
|
| 34 |
+
- π§± **Modular Tools**: Easily extendable with tools like web search, calculator, and custom APIs
|
| 35 |
+
|
| 36 |
+
## π¦ Stack
|
| 37 |
+
|
| 38 |
+
- **Frontend:** Gradio
|
| 39 |
+
- **Agent Framework:** LangChain (ReAct agent)
|
| 40 |
+
- **Vector DB:** Qdrant
|
| 41 |
+
- **LLM:** Gemini
|
| 42 |
+
- **Embedding Model:** BAAI/bge-large-en-v1.5
|
| 43 |
+
|
| 44 |
+
## π Example Queries
|
| 45 |
+
|
| 46 |
+
Try asking questions like:
|
| 47 |
+
|
| 48 |
+
1. **"What is LangChain and how is it different from LlamaIndex?"**
|
| 49 |
+
2. **"Who is the CEO of OpenAI and when was the company founded?"**
|
| 50 |
+
3. **"What is 245 * 92?"**
|
| 51 |
+
|
| 52 |
+
π‘ The chatbot decides whether to use RAG or call tools like calculator or web search automatically!
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
## π GitHub Repository
|
| 56 |
+
|
| 57 |
+
You can explore the full source code, Docker setup, and implementation details on GitHub:
|
| 58 |
+
|
| 59 |
+
[π RAGent Chatbot](https://github.com/shafiqul-islam-sumon/ragent-chatbot)
|
| 60 |
+
|
| 61 |
+
|
| 62 |
+
## π Try It Live
|
| 63 |
+
|
| 64 |
+
> Click **Open in Spaces** to launch the app
|
| 65 |
+
Or visit: [https://huggingface.co/spaces/shafiqul1357/rag-fashion-recommendation](https://huggingface.co/spaces/shafiqul1357/rag-fashion-recommendation)
|