Spaces:
Sleeping
Sleeping
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,12 +1,48 @@
|
|
| 1 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
|
| 3 |
This Space lets you upload company documents and/or crawl your website, build a vector index, and ask questions answered by a DeepSeek model via the Hugging Face Inference API.
|
| 4 |
|
| 5 |
## π Quick Start
|
| 6 |
1. Create a **new Space** β **Gradio** β Python.
|
| 7 |
2. Upload `app.py`, `requirements.txt`, `README.md`.
|
| 8 |
-
3. Add your Hugging Face token in Space **Settings β
|
| 9 |
4. Enter model ID (your fine-tuned DeepSeek model) in the UI.
|
| 10 |
5. Upload docs or crawl your site β build index β chat!
|
| 11 |
|
| 12 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
title: DeepSeek RAG Chatbot
|
| 3 |
+
emoji: π
|
| 4 |
+
colorFrom: indigo
|
| 5 |
+
colorTo: red
|
| 6 |
+
sdk: gradio
|
| 7 |
+
sdk_version: "4.44.0"
|
| 8 |
+
app_file: app.py
|
| 9 |
+
pinned: false
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# DeepSeek RAG β Chat With Your Website & Company Docs
|
| 13 |
|
| 14 |
This Space lets you upload company documents and/or crawl your website, build a vector index, and ask questions answered by a DeepSeek model via the Hugging Face Inference API.
|
| 15 |
|
| 16 |
## π Quick Start
|
| 17 |
1. Create a **new Space** β **Gradio** β Python.
|
| 18 |
2. Upload `app.py`, `requirements.txt`, `README.md`.
|
| 19 |
+
3. Add your Hugging Face token in Space **Settings β Variables and secrets**.
|
| 20 |
4. Enter model ID (your fine-tuned DeepSeek model) in the UI.
|
| 21 |
5. Upload docs or crawl your site β build index β chat!
|
| 22 |
|
| 23 |
---
|
| 24 |
+
|
| 25 |
+
## π§ How It Works
|
| 26 |
+
- Documents are split into chunks, embedded with `sentence-transformers/all-MiniLM-L6-v2`, and stored in a **FAISS** vector index.
|
| 27 |
+
- At query time, the top-k chunks are retrieved and sent with your question to a **DeepSeek LLM** via the Hugging Face Inference API.
|
| 28 |
+
|
| 29 |
+
---
|
| 30 |
+
|
| 31 |
+
## π¦ Dependencies
|
| 32 |
+
Managed by `requirements.txt`:
|
| 33 |
+
- gradio
|
| 34 |
+
- faiss-cpu
|
| 35 |
+
- sentence-transformers
|
| 36 |
+
- pypdf2
|
| 37 |
+
- docx2txt
|
| 38 |
+
- beautifulsoup4
|
| 39 |
+
- requests
|
| 40 |
+
- huggingface_hub
|
| 41 |
+
- numpy
|
| 42 |
+
|
| 43 |
+
---
|
| 44 |
+
|
| 45 |
+
## π Notes
|
| 46 |
+
- Keep your Hugging Face token private.
|
| 47 |
+
- Uploaded files are processed inside the Space (no external upload).
|
| 48 |
+
- Crawling is shallow and same-domain only by default.
|