metadata

license: mit
pipeline_tag: image-classification
tags:
  - image-classification
  - timm
  - transformers
  - detection
  - deepfake
  - forensics
  - deepfake_detection
  - community
  - opensight
base_model:
  - timm/vit_small_patch16_384.augreg_in21k_ft_in1k
library_name: transformers

Trained on 2.7M samples across 4,803 generators (see Training Data)

Uploaded for community validation as part of OpenSight - An upcoming open-source framework for adaptive deepfake detection.

Project OpenSight HF Spaces coming soon with an eval playground and eventually a leaderboard. Preview:

Model Details

Model Description

Vision Transformer (ViT) model trained on the largest dataset to-date for detecting AI-generated images in forensic applications.

Developed by: Jeongsoo Park and Andrew Owens, University of Michigan
Model type: Vision Transformer (ViT-Small)
License: MIT (compatible with CreativeML OpenRAIL-M referenced in [2411.04125v1.pdf])
Finetuned from: timm/vit_small_patch16_384.augreg_in21k_ft_in1k
Adapted for HF inference compatibility by AI Without Borders.

HF Space will be open sourced shortly showcasing various ways to run ultra-fast inference. Make sure to follow us for updates, as we will be releasing a slew of projects in the coming weeks.

Training Details

Training Data

2.7mil images from 15+ generators, 4600+ models
Over 1.15TB worth of images

Training Hyperparameters

Framework: PyTorch 2.0
Precision: bf16 mixed
Optimizer: AdamW (lr=5e-5)
Epochs: 10
Batch Size: 32

Evaluation

Unverified Testing Results

Only unverified because we currently lack resources to evaluate a dataset over 1.4T large.

Metric	Value
Accuracy	97.2%
F1 Score	0.968
AUC-ROC	0.992
FP Rate	2.1%

Re-sampled and refined dataset

Coming soon™

Citation

BibTeX:

@misc{park2024communityforensics,
    title={Community Forensics: Using Thousands of Generators to Train Fake Image Detectors}, 
    author={Jeongsoo Park and Andrew Owens},
    year={2024},
    eprint={2411.04125},
    archivePrefix={arXiv},
    primaryClass={cs.CV},
    url={https://arxiv.org/abs/2411.04125}, 
}

aiwithoutborders-xyz
/

OpenSight-CommunityForensics-Deepfake-ViT