List your company
ResearchFormerly WizWand

Audio Event Tagging on AudioSet AS-2M (full)

50.2mAPSSLAM
39.2842.11544.9547.785Jul 4, 2024Sep 22, 2024Dec 11, 2024Mar 2, 2025May 21, 2025Aug 9, 2025Oct 29, 2025
Updated 8mo ago

Evaluation results

MethodLinks
SSLAM#Param=88M, Pre-training Data=AS2025.0650.2
CEDType=Reference model, Evaluation Protocol=Fine-tuning2025.0350
SPEARs+a XLarge# Params=600M, Pre-train data=197k2025.1050
SPEARa Large# Params=327M, Pre-train data=13k2025.1049.8
ATST-C2FType=Reference model, Evaluation Protocol=Fine-tuning2025.0349.7
SPEARa Large# Params=327M, Pre-train data=5k2025.1049.7
SPEARs+a Large# Params=327M, Pre-train data=97k2025.1049.6
SPEARa Base# Params=94M, Pre-train data=5k2025.1049.3
SPEARa Base# Params=94M, Pre-train data=13k2025.1049.3
M2D-CLAP2025 stage1Type=Audio-Language model, Evaluation Protocol=Fine-tuning2025.0349
DyMN-LType=Reference model, Evaluation Protocol=Fine-tuning2025.0349
DASS-MediumParams=49M, Pretrain=IN SL, Teacher ensemble=AST + HTS-AT2024.0748.9
BEATSiter3+#Param=90M, Pre-training Data=AS, Supervised Fine-tuning=Included2025.0648.6
A-JEPA#Param=86M, Pre-training Data=AS2025.0648.6
EAT#Param=88M, Pre-training Data=AS2025.0648.6
BEATs(iter3)Params=90M, Pretrain=SSL2024.0748.6
EATParams=88M, Pretrain=SSL2024.0748.6
DASS-SmallParams=30M, Pretrain=IN SL, Teacher ensemble=AST + HTS-AT2024.0748.6
BEATSiter3+Type=Audio model, Evaluation Protocol=Fine-tuning2025.0348.6
DTF-AT Full TFType=Audio model, Evaluation Protocol=Fine-tuning2025.0348.6
BEATs# Params=90M, Pre-train data=5k2025.1048.6
EAT# Params=88M, Pre-train data=5k2025.1048.6
SPEARs+a Base# Params=94M, Pre-train data=97k2025.1048.6
M2D-CLAP2024 stage1Type=Ablation, Evaluation Protocol=Fine-tuning2025.0348.5
SupMAM-CLAPType=Reference model, Evaluation Protocol=Fine-tuning2025.0348.5
BEATSiter3#Param=90M, Pre-training Data=AS2025.0648
ASIT#Param=86M, Pre-training Data=AS2025.0648
BEATSiter3Type=Audio model, Evaluation Protocol=Fine-tuning2025.0348
ATST-FrameType=Audio model, Evaluation Protocol=Fine-tuning2025.0348
ATST Frame# Params=86M, Pre-train data=5k2025.1048
M2DType=Baseline, Evaluation Protocol=Fine-tuning2025.0347.9
DASS-MediumParams=49M, Pretrain=IN SL2024.0747.6
Audio-MAE#Param=86M, Pre-training Data=AS2025.0647.3
Audio-MAEParams=86M, Pretrain=SSL2024.0747.3
DASS-SmallParams=30M, Pretrain=IN SL2024.0747.2
MaskSpec#Param=86M, Pre-training Data=AS2025.0647.1
HTS-ATParams=31M, Pretrain=IN SL2024.0747.1
PaSSTPretrain=IN SL2024.0747.1
HTS-ATType=Audio model, Evaluation Protocol=Fine-tuning2025.0347.1
ASTParams=87M, Pretrain=IN SL2024.0745.9
ATST-ClipType=Audio model, Evaluation Protocol=Fine-tuning2025.0345.2
Audio MambaParams=40M, Pretrain=IN SL2024.0744
SPEARs Large# Params=327M, Pre-train data=84k2025.1043.9
SPEARs Base# Params=94M, Pre-train data=84k2025.1043.6
AuMParams=26M, Pretrain=IN SL2024.0739.7