---
license: mit
tags:
- deepfake-detection
- video-analysis
- temporal-consistency
---

# ProofGuard Temporal Detector

This model detects deepfakes in videos by analyzing temporal consistency across frames.

## Model Details
- **Architecture**: TemporalConsistencyDetector
- **Input**: Video sequences (16 frames)
- **Output**: Binary classification (Real/Fake)

## Usage
```python
import torch

# Load the model checkpoint
checkpoint = torch.load("pytorch_model.bin", map_location="cpu")

# Access model weights
model_state_dict = checkpoint['model_state_dict']

# The model analyzes temporal patterns in video sequences
# to detect inconsistencies that indicate deepfakes
```

## Components

- Frame feature extractor (EfficientNet backbone)  
- Bidirectional LSTM for temporal modeling  
- Temporal attention mechanism  
- Binary classifier

Trained on diverse video datasets containing both authentic and deepfake content.

## Author
Onome Akpobaro