List your company

Audio-to-Text Retrieval on Clotho

26.5R@1Cacophony
-0.9566.17213.320.428Dec 11, 2025Jan 7, 2026Feb 4, 2026Mar 4, 2026Mar 31, 2026Apr 28, 2026May 26, 2026

Evaluation results

MethodLinks
CacophonyYear=2024, Setting=ZS2026.0526.5————
MGA-CLAPYear=2024, Setting=ZS2026.0526.5————
CacophonyYear=2024, Setting=ZS2026.0526.5————
MGA-CLAPYear=2024, Setting=ZS2026.0526.5————
WavCaps-CNN14Year=2023, Setting=ZS2026.0525.9————
WavCaps-CNN14Year=2023, Setting=ZS2026.0525.9————
LAION-CLAP fusionYear=2023, Setting=ZS2026.0525.7————
LAION-CLAP fusionYear=2023, Setting=ZS2026.0525.7————
FLAP fusionYear=2024, Setting=ZS2026.0525.5————
FLAP fusionYear=2024, Setting=ZS2026.0525.5————
M2D-CLAPYear=2025, Setting=ZS2026.0525————
M2D-CLAPYear=2025, Setting=ZS2026.0525————
WavCaps-HTSATYear=2023, Setting=ZS2026.0523.4————
WavCaps-HTSATYear=2023, Setting=ZS2026.0523.4————
Microsoft-CLAPModality=a2t2026.0423.247.557.6—15.4
GLAPYear=2025, Setting=ZS2026.0521.8————
GLAPYear=2025, Setting=ZS2026.0521.8————
WavCaps-CNN14Year=2023, Setting=ZS2026.0521.2————
MGA-CLAPYear=2024, Setting=ZS2026.0520.8————
FLAP fusionYear=2024, Setting=ZS2026.0520.3————
CacophonyYear=2024, Setting=ZS2026.0520.2————
M2D-CLAPYear=2025, Setting=ZS2026.0520.1————
MS-CLAPYear=2023, Setting=ZS2026.0520————
MS-CLAPYear=2023, Setting=ZS2026.0520————
WavCaps-HTSATYear=2023, Setting=ZS2026.0519.5————
GLAPYear=2025, Setting=ZS2026.0519.4————
OmniRetriever-7BYear=2026, Setting=ZS2026.0519.14————
ProposedModality=a2t, audio encoder=HTSAT-tiny, text encoder=RoBERTa-large2026.0418.348.259.1—16.2
LAION-CLAP fusionYear=2023, Setting=ZS2026.0518.2————
LAION-CLAPModality=a2t2026.0417.53745.5—15.5
MS-CLAPYear=2023, Setting=ZS2026.0516.7————
LanguageBind (CLIP-H/14)Year=2024, Setting=ZS2026.0516.7————
OmniRetriever-7BYear=2026, Setting=ZS2026.0516.1————
OmniRetriever-7BYear=2026, Setting=ZS2026.0516.08————
MMTYear=2022, Setting=FT2026.057————
MMTYear=2022, Setting=FT2026.057————
MMTYear=2022, Setting=FT2026.056.7————
ImageBind (ViT-H)Year=2023, Setting=ZS2026.056————
Gemini Embedding 2 (closed)Year=2026, Setting=ZS2026.055.19————
SGD-CLIPTime=347.48s2026.043.7312.4419.23——
Omni-Embed-NemotronYear=2025, Setting=ZS2026.053.5————
UniConTime=13.45s2026.043.3513.1119.43——
Gemini Embedding 2 (closed)Year=2026, Setting=ZS2026.051.34————
Gemini Embedding 2Year=2026, Setting=ZS2026.051.3————
ASK+Interaction Strategy=Local, Knowledge Source=WavCaps2025.120.1950.4140.545——
ASK*Interaction Strategy=Local, Knowledge Source=Original training set2025.120.1920.4160.545——
ASK†Interaction Strategy=Local, Knowledge Source=Gemini-annotated training set2025.120.1910.4190.539——
Xie et al., 2024Interaction Strategy=Local2025.120.1810.4020.534——
No alignment2026.040.10.480.96——
Reference Model (Mei et al., 2024)2025.09———27.1—
WAVE 7BModel Size=7B2025.09———24.5—