# Finance Classifier Model

This directory contains the fine-tuned mBERT model for binary financial conversation classification.

## Model Files

The model directory should contain:
- `config.json` - Model configuration
- `tokenizer_config.json` - Tokenizer configuration  
- `special_tokens_map.json` - Special tokens mapping
- `pytorch_model.bin` - Trained model weights (generated by training)

## Training

To generate the trained model, run:

```bash
cd nlp/
python train_classifier.py
```

This will:
1. Load training data from `../classifier_training.json`
2. Fine-tune bert-base-multilingual-cased on financial vs non-financial classification
3. Save the trained model to this directory

## Model Details

- **Base Model**: bert-base-multilingual-cased
- **Task**: Binary Classification (financial: 1, non-financial: 0)
- **Input**: Text sentences
- **Languages**: Multilingual support
- **Training File**: `classifier_training.json`

## Usage

```python
from nlp.classifier import FinanceClassifier

clf = FinanceClassifier()
result = clf.predict("Loan lena chahiye")
print(result)  # {'prediction': 'financial', 'confidence': 0.95}
```