Upload README.md with huggingface_hub

README.md

@@ -1,12 +1,58 @@
----
-title: Atlas 1 Demo
-emoji: 🚀
-colorFrom: gray
-colorTo: gray
-sdk: gradio
-sdk_version: 6.2.0
-app_file: app.py
-pinned: false
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+---
+title: Atlas 1 Demo
+emoji: 🧪
+colorFrom: blue
+colorTo: purple
+sdk: gradio
+sdk_version: 4.0.0
+app_file: app.py
+pinned: false
+license: mit
+tags:
+- molecular-property-prediction
+- chemistry
+- graph-neural-network
+- 3d-visualization
+---
+
+# Atlas 1: Molecular Property Prediction Demo
+
+Interactive demo for **Atlas 1**, a state-of-the-art Geometric GNN that predicts molecular properties (HOMO energy) from 3D structures.
+
+## Features
+
+- 🔬 **2D Molecular Structures** - Clean chemical diagrams
+- 🌐 **3D Interactive View** - Rotatable molecules with proper atom colors
+- ⚡ **Real-time Predictions** - Instant HOMO energy predictions
+- 📊 **Model Performance** - RMSE: 0.109 eV, MAE: 0.061 eV, Pearson r: 0.994
+- 🎨 **Professional UI** - Modern Gradio interface with examples
+
+## How to Use
+
+1. Enter a SMILES string (e.g., `C1=CC=CC=C1` for benzene)
+2. Click "Predict HOMO Energy"
+3. View the 2D structure, 3D interactive molecule, and prediction results
+
+## Example Molecules
+
+- **Methane**: `C`
+- **Benzene**: `C1=CC=CC=C1`
+- **Caffeine**: `CN1C=NC2=C1C(=O)N(C(=O)N2C)C`
+- **Ethanol**: `CCO`
+
+## Model Details
+
+- **Architecture**: CFConv + SE(3)-equivariant features
+- **Parameters**: 2.5M
+- **Training**: QM9 dataset (130k molecules)
+- **Performance**: 2.2× better than baseline
+
+## Links
+
+- **Model**: [Reverb/atlas-1-molecular-property-prediction](https://huggingface.co/Reverb/atlas-1-molecular-property-prediction)
+- **GitHub**: Atlas Project
+- **Paper**: Coming soon
+
+---
+
+**Part of the Atlas Series**: From properties to structures