# Email Classification Model (Simple Version)

A dual-head transformer model for classifying healthcare emails into categories and subcategories.

## Model Details
- **Base Model**: distilbert-base-uncased
- **Categories**: 6
- **Subcategories**: 14

## Categories
appointments, denials, eligibility, other, patient_balance, submission

## Usage

```python
import torch
from transformers import AutoModel, AutoTokenizer

# Download the inference script
# wget https://huggingface.co/commure-smislam/email-classification-simple/resolve/main/inference.py

from inference import EmailClassifierInference

# Load model
classifier = EmailClassifierInference("./")

# Predict
result = classifier.predict("Patient appointment confirmation for tomorrow")
print(result)
```

## Expected Output
```json
{
  "text": "Patient appointment confirmation for tomorrow",
  "category": {"label": "appointments", "confidence": 0.95},
  "subcategory": {"label": "appointments.confirmation", "confidence": 0.92}
}
```

## Files
- `backbone/`: Base transformer model
- `tokenizer/`: Tokenizer files  
- `classification_heads.pt`: Classification layer weights
- `inference.py`: Inference script