Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Wild

Benchmarks

Task NameDataset NameSOTA ResultTrend
Single Image Reflection RemovalWild 55 images (test)
PSNR29.54
65
Generative SteganographyWILD
Entropy (bit/token)2.1539
21
Image Reflection RemovalWild
PSNR26.11
20
Speech Quality AssessmentWild
MOS0.85
18
Audio Anti-spoofingWild
EER0.0121
12
Transmission Layer SeparationWild 55 images (test)
PSNR27.7
12
Source AttributionWILD Closed-set (Raw)
Adobe Firefly98.1
10
Speech-driven talking face generationWild Dataset
LSE-D8.89
8
Single Image Reflection RemovalWild (test)
PSNR26.24
8
Open-set rejectionWILD Open-set
AUC (%)100
7
Unknown Generator DiscoveryWILD
Novel NMI0.9
5
Code Injection detection and exploitationWild npm registry (crawled)
Detected Count20
4
Command Injection detection and exploitationWild npm registry (crawled)
Detected Count28
4
Prototype Pollution detection and exploitationWild npm registry (crawled)
Detected Count20
3
Path Traversal detection and exploitationWild npm registry (crawled)
Detected Count44
3
Multi-condition Human Motion SynthesisWild
Beat Alignment69
1
Showing 16 of 16 rows