tomas-gajarsky/facetorch-deepfake-efficientnet-b7

Deepfake Detection

Exported PyTorch model (.pt2) for use with facetorch.

Model Details

Task	Deepfake Detection
Architecture	EfficientNet-B7
Format	torch.export (.pt2) — no model source code needed
Dynamic shapes	Batch dimension is dynamic (1-64)
Input	380x380 face crop
Output	1-dim sigmoid score (Real vs Fake)

Original Work

This model is based on selimsef/dfdc_deepfake_challenge. Weights converted and exported by facetorch.

Usage

import torch

# Load — no model class needed
ep = torch.export.load("model.pt2")
model = ep.module()

# Inference
x = torch.randn(1, 3, 224, 224)  # adjust size per model
output = model(x)

Or via facetorch:

from facetorch import FaceAnalyzer
from omegaconf import OmegaConf

cfg = OmegaConf.load("conf/config.yaml")
analyzer = FaceAnalyzer(cfg.analyzer)
response = analyzer.run(path_image="face.jpg")