| --- |
| title: PatchWise |
| emoji: 🚀 |
| colorFrom: blue |
| colorTo: purple |
| sdk: streamlit |
| sdk_version: "1.45.1" |
| python_version: "3.9" |
| app_file: app.py |
| pinned: false |
| --- |
| |
| # 🚀 PatchWise |
|
|
| PatchWise is a Reinforcement Learning-based Adaptive Vision Transformer designed for efficient edge AI inference. |
|
|
| ## Features |
| - Dynamic patch routing |
| - Sparse transformer inference |
| - A3C reinforcement learning policy |
| - Adaptive compute allocation |
| - Patch importance visualization |
| - Edge deployment ready |
|
|
| ## Model Highlights |
| - ~40% FLOPs reduction |
| - Real-time adaptive routing |
| - NVIDIA Jetson AGX Orin deployment |
| - Interactive Streamlit demo |
|
|
| ## How It Works |
|
|
| Image → RL Policy → Sparse Vision Transformer → Adaptive Inference |
