README.md

---
title: Multimodal A/B Test Predictor
emoji: 🚀
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: "4.44.0"
app_file: app.py
pinned: false
---

# 🚀 Multimodal A/B Test Predictor

## Overview
Advanced A/B testing outcome predictor using multimodal AI analysis combining:
- 🖼️ **Image Analysis**: Visual features from control & variant images
- 📝 **OCR Text Extraction**: Automatically extracts and analyzes text from images  
- 📊 **Categorical Features**: Business context provided via API (industry, page type, etc.)
- 🎯 **Confidence Scores**: Based on training data statistics and historical accuracy

## 🎯 Direct Input Architecture

### **Image + Categorical Data**
- Accepts control and variant images directly via API
- Requires categorical inputs: Business Model, Customer Type, Conversion Type, Industry, Page Type
- Fast, efficient predictions without external API dependencies
- All processing happens locally on GPU

## 🎯 Features

### Direct Prediction with Categorical Data
- Upload control & variant images
- Provide categorical business context data
- Fast prediction with comprehensive confidence analysis

### Enhanced Results
- **Winner Prediction**: Variant vs Control with probability
- **Model Confidence**: Accuracy percentage from training data
- **Training Data Count**: Number of samples model trained on for this category
- **Historical Win/Loss**: Real A/B test outcome statistics for this category
- **Confidence Source**: Industry + Page Type combination used for scoring

## 🔧 Setup

### Model Files
- `model/multimodal_gated_model_2.7_GGG.pth`: Enhanced multimodal model (789MB)
- `model/multimodal_cat_mappings_GGG.json`: Category mappings

## 🚀 Technical Architecture

### Model: SupervisedSiameseMultimodal (GGG Enhanced)
- **Vision**: ViT (Vision Transformer) for image features
- **Text**: DistilBERT for OCR text processing
- **Fusion**: Gated fusion with directional features
- **Categories**: Embedding layers for categorical features
- **Architecture**: BatchNorm + Fusion Block + Enhanced Prediction Head

### Confidence Scoring
- Based on Industry + Page Type combinations
- Uses holdout statistics with average 160 samples per combination
- Much more reliable than full 5-feature combinations

## 📊 Performance
- **Multimodal Analysis**: Images + Text + Categories
- **GPU Accelerated**: Fast predictions (2-4 seconds average)
- **High Accuracy**: Enhanced GGG architecture with real training data
- **No External Dependencies**: All processing done locally

## 🎯 Use Cases
- **A/B Test Prediction**: Predict winners before running tests with provided context
- **Batch Processing**: Process multiple tests efficiently from CSV
- **Confidence Assessment**: Understand prediction reliability based on historical data
- **API Integration**: Easy integration with external systems

## 📡 API Usage

### Quick Start
```python
from gradio_client import Client

client = Client("SpiralyzeLLC/ABTestPredictor")
result = client.predict(
    "control.jpg", "variant.jpg",
    "SaaS", "B2B", "High-Intent Lead Gen",
    "B2B Software & Tech", "Awareness & Discovery",
    api_name="/predict_with_categorical_data"
)
```

### 📚 Documentation

- **[QUICK_START_GUIDE.md](QUICK_START_GUIDE.md)** - Get started in 3 steps
- **[API_DOCUMENTATION.md](API_DOCUMENTATION.md)** - Complete API reference with examples in Python, JavaScript, and cURL
- **[API_USAGE_UPDATED.md](API_USAGE_UPDATED.md)** - Technical details and migration notes

### 🌐 Web Interface

Use the interactive interface: **https://huggingface.co/spaces/SpiralyzeLLC/ABTestPredictor**

---

Built with ❤️ using Gradio, PyTorch, Transformers, and Hugging Face.