addd HF space config in readme

README.md

@@ -1,3 +1,14 @@
+---
+title: MNIST 2x2 Transformer Demo
+emoji: 🔢
+colorFrom: gray
+colorTo: blue
+sdk: gradio
+sdk_version: 5.29.0  # or your current gradio version
+app_file: app.py
+pinned: false
+---
+
 # Transformer MNIST 2×2 — Image-to-Sequence Prediction
 
 This project implements a minimal Transformer-based model that takes a 2×2 grid of MNIST digits as input and autoregressively predicts the corresponding 4-digit sequence. It serves as a practical deep dive into the inner workings of the Transformer architecture and basic multimodality concepts, combining vision (image patches) with language modeling (digit sequences).