Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,48 @@
|
|
| 1 |
---
|
|
|
|
|
|
|
| 2 |
license: mit
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
license: mit
|
| 5 |
+
tags:
|
| 6 |
+
- rag
|
| 7 |
+
- finance
|
| 8 |
+
- sec-filings
|
| 9 |
+
- retrieval-augmented-generation
|
| 10 |
+
- chromadb
|
| 11 |
+
- bm25
|
| 12 |
+
- gradio
|
| 13 |
+
pipeline_tag: text-generation
|
| 14 |
---
|
| 15 |
+
|
| 16 |
+
# Financial RAG System
|
| 17 |
+
|
| 18 |
+
This repository documents a Financial Retrieval-Augmented Generation system for answering questions over SEC 10-K and 10-Q filings.
|
| 19 |
+
|
| 20 |
+
Live demo: https://huggingface.co/spaces/anasxs/financial-rag
|
| 21 |
+
|
| 22 |
+
GitHub: https://github.com/Anassbzdd/financial-RAG
|
| 23 |
+
|
| 24 |
+
## System Overview
|
| 25 |
+
|
| 26 |
+
This is not a fine-tuned language model. It is a RAG system that combines document parsing, chunking, vector retrieval, keyword retrieval, reranking, and grounded answer generation.
|
| 27 |
+
|
| 28 |
+
The system answers questions using indexed SEC filings from:
|
| 29 |
+
|
| 30 |
+
- Apple
|
| 31 |
+
- Amazon
|
| 32 |
+
- Alphabet
|
| 33 |
+
- Berkshire Hathaway
|
| 34 |
+
- Johnson and Johnson
|
| 35 |
+
|
| 36 |
+
## Architecture
|
| 37 |
+
|
| 38 |
+
```text
|
| 39 |
+
SEC filings
|
| 40 |
+
-> LlamaParse markdown extraction
|
| 41 |
+
-> Recursive text chunking
|
| 42 |
+
-> BGE embeddings
|
| 43 |
+
-> ChromaDB vector search
|
| 44 |
+
-> BM25 keyword search
|
| 45 |
+
-> Reciprocal Rank Fusion
|
| 46 |
+
-> Cross-encoder reranking
|
| 47 |
+
-> Groq LLM generation
|
| 48 |
+
-> Gradio UI
|