Spaces:
Running
Running
Update README.md
Browse files
README.md
CHANGED
|
@@ -3,54 +3,9 @@ title: SupraChat
|
|
| 3 |
emoji: 🤖
|
| 4 |
colorFrom: blue
|
| 5 |
colorTo: gray
|
| 6 |
-
sdk:
|
| 7 |
-
sdk_version:
|
| 8 |
app_file: app.py
|
| 9 |
pinned: false
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
# SupraChat – ChatGPT‑like Interface for SupraLabs Models
|
| 13 |
-
|
| 14 |
-
A clean, modern chat interface built with Gradio, supporting all text‑based models from [SupraLabs](https://huggingface.co/SupraLabs).
|
| 15 |
-
Conversation history is stored temporarily in RAM and is cleared when you leave the page.
|
| 16 |
-
|
| 17 |
-
## Supported Models
|
| 18 |
-
|
| 19 |
-
- **Supra-50M-Instruct** – 50M instruction‑tuned model for general dialogue.
|
| 20 |
-
- **Supra-50M-Reasoning** – 50M reasoning model that outputs a thought process and a solution.
|
| 21 |
-
- **Supra-1.5-50M-Instruct-exp** – Experimental 50M model with 5K context length.
|
| 22 |
-
- **Supra-50M-Base** – Pure base model (no instruction tuning).
|
| 23 |
-
- **StorySupra-10M** – 10M model specialised for story generation.
|
| 24 |
-
- **Supra-Mini-v5-8M** – Ultra‑small 8M model for fast experimentation.
|
| 25 |
-
|
| 26 |
-
## Features
|
| 27 |
-
|
| 28 |
-
- 💬 ChatGPT‑style conversation interface
|
| 29 |
-
- 🔄 Switch between any supported model on the fly
|
| 30 |
-
- 📝 Automatic conversation title generation using `Supra-Title-Flan-85M`
|
| 31 |
-
- 💾 History stored in memory (cleared on page close)
|
| 32 |
-
- 🎨 Clean, responsive UI with adjustable temperature and token length
|
| 33 |
-
|
| 34 |
-
## Usage
|
| 35 |
-
|
| 36 |
-
1. Select a model from the dropdown.
|
| 37 |
-
2. Adjust temperature (creativity) and maximum output length.
|
| 38 |
-
3. Type a message and press Enter or click Send.
|
| 39 |
-
4. The title is auto‑generated from the first message.
|
| 40 |
-
5. To start a new conversation, click “Clear Chat”.
|
| 41 |
-
|
| 42 |
-
## Technical Details
|
| 43 |
-
|
| 44 |
-
- The app loads models on demand and caches them for subsequent requests.
|
| 45 |
-
- For reasoning models, the prompt is automatically prefixed with `<|begin_of_thought|>` to trigger the thought process.
|
| 46 |
-
- All dependencies are listed in `requirements.txt`.
|
| 47 |
-
|
| 48 |
-
## Deployment on Hugging Face Spaces
|
| 49 |
-
|
| 50 |
-
1. Create a new Space with Gradio SDK.
|
| 51 |
-
2. Upload `app.py`, `requirements.txt`, and this `README.md`.
|
| 52 |
-
3. The Space will install dependencies and launch the app.
|
| 53 |
-
|
| 54 |
-
## License
|
| 55 |
-
|
| 56 |
-
This project is for demonstration purposes. The underlying models are released by SupraLabs under their respective licenses.
|
|
|
|
| 3 |
emoji: 🤖
|
| 4 |
colorFrom: blue
|
| 5 |
colorTo: gray
|
| 6 |
+
sdk: docker
|
| 7 |
+
sdk_version: 6.19.0
|
| 8 |
app_file: app.py
|
| 9 |
pinned: false
|
| 10 |
+
short_description: Inference UI for all SupraLabs models
|
| 11 |
+
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|