Views
No views yet
| File | Source | Task |
|---|---|---|
generative_detector.onnx | Community Forensics ViT-S/384 (arXiv:2411.04125, MIT) — exported from the original timm checkpoint model_v11_ViT_384_base_ckpt.pt | AI-generated image detection. Input (N,3,384,384) f32, resize-440/center-crop-384, CLIP mean/std. Output (N,1) logit; sigmoid = P(AI-generated). |
audio_deepfake_detector.onnx | MelodyMachine/Deepfake-audio-detection-V2 (wav2vec2) | Synthesized-voice detection. Input (N,samples) f32, 16 kHz mono, per-chunk zero-mean/unit-var. Output (N,2) logits: [fake, real]. |
training/export_hf_models.py