CommunityForensics-DeepfakeDet-ViT is an image classification model hosted on Hugging Face. It forms part of a community-driven effort in media forensics and is designed to classify images as real or manipulated.
The model is built on the Vision Transformer architecture and carries the tags detection, deepfake, forensics, and deepfake_detection. It was developed under the Borderless R&D organization and is associated with the arxiv paper 2411.04125. The implementation uses the timm library and is provided in Safetensors format.
It can be used through the Transformers library via a high-level pipeline for image classification or by directly loading the AutoImageProcessor and AutoModelForImageClassification components. Example code demonstrates loading the model and running inference on an image URL. The model card indicates availability for deployment through Hugging Face inference providers, notebooks, and local applications.
The project is released under the MIT license and is openly accessible on the Hugging Face platform. It belongs to the class of transformer-based image classification models focused on deepfake detection.
CommunityForensics DeepfakeDet ViT sits in PulseGate's Other AI category. Automatically detecting AI-generated or manipulated images and deepfakes in digital media. It is built as an open-source project for developers. CommunityForensics DeepfakeDet ViT is open source under the MIT license. The product ships for the web and API.
buildborderless builds and maintains CommunityForensics DeepfakeDet ViT, and the product first shipped in 2025. Development happens publicly on GitHub with 46 stars. Key capabilities include Image Classification, Deepfake Detection, and forensics.
Latest indexed changes and source events
buildborderless/CommunityForensics-DeepfakeDet-ViT verified by the PulseGate indexer
Other apps tracked under the same category.
