Spaces:
Build error
Build error
Update README.md
Browse files
README.md
CHANGED
@@ -10,4 +10,29 @@ pinned: false
|
|
10 |
short_description: RAG System with Phi-2
|
11 |
---
|
12 |
|
13 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
short_description: RAG System with Phi-2
|
11 |
---
|
12 |
|
13 |
+
# PDF RAG System with Phi-2
|
14 |
+
|
15 |
+
This application allows you to upload PDF documents and ask questions about their content. It uses a Retrieval-Augmented Generation (RAG) pipeline with the Phi-2 language model.
|
16 |
+
|
17 |
+
## Features
|
18 |
+
|
19 |
+
- PDF document upload and processing
|
20 |
+
- Text and table extraction from PDFs
|
21 |
+
- Hybrid search combining vector embeddings and BM25
|
22 |
+
- Query expansion for better search results
|
23 |
+
- Streaming response generation
|
24 |
+
|
25 |
+
## How to Use
|
26 |
+
|
27 |
+
1. Upload your PDF files
|
28 |
+
2. Process the documents
|
29 |
+
3. Ask questions about the content
|
30 |
+
4. Adjust retrieval and generation parameters as needed
|
31 |
+
|
32 |
+
## Technical Details
|
33 |
+
|
34 |
+
- Uses Phi-2 language model via llama.cpp
|
35 |
+
- Sentence Transformers for document embeddings
|
36 |
+
- FAISS for vector search
|
37 |
+
- BM25 for keyword search
|
38 |
+
- Hybrid retrieval with configurable weighting
|