List your company
ResearchFormerly WizWand

Audio Classification on AudioSet 2M

50.5mAPMiCo
18.2626.633543.37Dec 18, 2022Jul 13, 2023Feb 5, 2024Aug 30, 2024Mar 25, 2025Oct 18, 2025May 14, 2026
Updated 4mo ago

Evaluation results

MethodLinks
MiCoBackbone=ViT-g2024.0650.5
SSLAM# Params=88M, Data=AS-2M, Source=Reported, Evaluation Protocol=Fine-tuning (FT)2026.0250.2
SSLAMBackbone=ViT-B/16, Pre-training Data=AS, Pre-training Parameters=93M, Fine-tuning Parameters=88M, Evaluation Protocol=fine-tuning2026.0550.2
AudioMosaicBackbone=ViT-B/16, Pre-training Data=AS, Pre-training Parameters=86M, Fine-tuning Parameters=86M, Evaluation Protocol=fine-tuning2026.0550.2
CEDFine-tuning=true2024.0650
ATST-C2F# Param=86M, Pre-training data=AS, knowledge distillation=true2023.0649.7
ATST-C2FFine-tuning=true2024.0649.7
BEATS_iter3+#Param=90M, Pre-training Data=AS2022.1248.6
BEATSiter3+# Param=90M, Pre-training data=AS, knowledge distillation=true2023.0648.6
BEATS_iter3+# Params=90M, Fine-tuning=true2024.0448.6
BEATsFine-tuning=true, # Params=90M, Iteration=3+2024.0448.6
BEATs iter3+Fine-tuning=true2024.0648.6
Absolute SOTA2024.0648.6
EAT# Params=88M, Data=AS-2M, Source=Reported, Evaluation Protocol=Fine-tuning (FT)2026.0248.6
A-JEPABackbone=ViT-B/16, Pre-training Data=AS, Pre-training Parameters=354M, Fine-tuning Parameters=86M, Evaluation Protocol=fine-tuning2026.0548.6
EATBackbone=ViT-B/16, Pre-training Data=AS, Pre-training Parameters=93M, Fine-tuning Parameters=88M, Evaluation Protocol=fine-tuning2026.0548.6
BAT# Params=91M, Data=AS-2M, Source=Our tuning/probing, Evaluation Protocol=Fine-tuning (FT)2026.0248.55
M2D-AS/0.7# Params=86M, Masking ratio=0.7, Fine-tuning=true2024.0448.5
SupMAM-CLAPFine-tuning=true2024.0648.5
M2D-CLAPFine-tuning=true, masking ratio=0.72024.0648.5
BEATS_iter2#Param=90M, Pre-training Data=AS2022.1248.1
BEATS_iter3#Param=90M, Pre-training Data=AS2022.1248
BEATSiter3# Param=90M, Pre-training data=AS2023.0648
ATST-Frame# Param=86M, Pre-training data=AS2023.0648
BEATsFine-tuning=true, # Params=90M, Iteration=32024.0448
BEATs iter3Fine-tuning=true2024.0648
ATST-FrameFine-tuning=true2024.0648
BEATsiter3Backbone=ViT-B/16, Pre-training Data=AS, Pre-training Parameters=182M, Fine-tuning Parameters=91M, Evaluation Protocol=fine-tuning2026.0548
BEATS_iter1#Param=90M, Pre-training Data=AS2022.1247.9
M2D/0.7# Params=86M, Masking ratio=0.7, Fine-tuning=true2024.0447.9
M2DFine-tuning=true, # Params=86M, Masking ratio=0.72024.0447.9
M2DFine-tuning=true, masking ratio=0.72024.0647.9
M2DFine-tuning=true, # Params=86M, Masking ratio=0.62024.0447.7
M2DFine-tuning=true, masking ratio=0.62024.0647.7
SSLAM# Params=88M, Data=AS-2M, Source=Our tuning/probing, Evaluation Protocol=Fine-tuning (FT)2026.0247.69
EAT# Params=88M, Data=AS-2M, Source=Our tuning/probing, Evaluation Protocol=Fine-tuning (FT)2026.0247.61
M2DFine-tuning=true, # Params=86M, Masking ratio=0.7, Target input strategy=all patches to target2024.0447.6
ASiTBackbone=ViT-B/16, Pre-training Data=AS, Pre-training Parameters=96M, Fine-tuning Parameters=86M, Evaluation Protocol=fine-tuning2026.0547.5
Audio-MAE Large#Param=304M, Pre-training Data=AS, Pre-training Type=Self-Supervised Pre-Training2022.1247.4
MSM-MAEFine-tuning=true, # Params=86M, Masking ratio=0.752024.0447.4
M2DFine-tuning=true, # Params=86M, Masking ratio=0.6, Target input strategy=all patches to target2024.0447.4
MSM-MAEFine-tuning=true, masking ratio=0.752024.0647.4
Audio-MAE#Param=86M, Pre-training Data=AS, Pre-training Type=Self-Supervised Pre-Training2022.1247.3
Audio-MAE (local)# Param=86M, Pre-training data=AS2023.0647.3
Audio-MAE (local)Fine-tuning=true, # Params=86M2024.0447.3
Audio-MAEBackbone=ViT-B/16, Pre-training Data=AS, Pre-training Parameters=137M, Fine-tuning Parameters=86M, Evaluation Protocol=fine-tuning2026.0547.3
PaSST#Param=86M, Pre-training Data=IN, Pre-training Type=Out-of-domain Supervised Pre-Training2022.1247.1
HTS-AT#Param=31M, Pre-training Data=IN, Pre-training Type=Out-of-domain Supervised Pre-Training2022.1247.1
PaSST#Param=86M, Pre-training Data=IN+AS, Pre-training Type=In-domain Supervised Pre-Training2022.1247.1
HTS-AT#Param=31M, Pre-training Data=IN+AS, Pre-training Type=In-domain Supervised Pre-Training2022.1247.1
MaskSpec#Param=86M, Pre-training Data=AS, Pre-training Type=Self-Supervised Pre-Training2022.1247.1
HTS-AT# Param=31M, Pre-training data=AS2023.0647.1
PassT# Param=86M, Pre-training data=AS2023.0647.1
KD-AST# Param=86M, Pre-training data=AS2023.0647.1
MaskSpec# Param=86M, Pre-training data=AS2023.0647.1
HTS-AT# Params=31M, Fine-tuning=true2024.0447.1
MaskSpecFine-tuning=true, # Params=86M2024.0447.1
PaSSTFine-tuning=true, # Params=86M2024.0447.1
HTS-ATFine-tuning=true, # Params=31M2024.0447.1
HTS-ATFine-tuning=true2024.0647.1
MaskSpecBackbone=ViT-B/16, Pre-training Data=AS, Pre-training Parameters=112M, Fine-tuning Parameters=86M, Evaluation Protocol=fine-tuning2026.0547.1
ATST-F2C# Param=86M, Pre-training data=AS, knowledge distillation=true2023.0646.8
AST#Param=86M, Pre-training Data=IN, Pre-training Type=Out-of-domain Supervised Pre-Training2022.1245.9
AST#Param=86M, Pre-training Data=IN+AS, Pre-training Type=In-domain Supervised Pre-Training2022.1245.9
AST# Param=86M, Pre-training data=AS2023.0645.9
AST (Single)# Params=86M, Fine-tuning=true2024.0445.9
ASTFine-tuning=true, # Params=86M, Version=Single2024.0445.9
ASTBackbone=ViT-B/16, Pre-training Data=IN, Pre-training Parameters=86M, Fine-tuning Parameters=86M, Evaluation Protocol=fine-tuning2026.0545.9
ATST-Clip# Param=86M, Pre-training data=AS2023.0645.2
ATST-ClipFine-tuning=true2024.0645.2
ERANN#Param=55M, Pre-training Type=No Pre-Training2022.1245
ERANNBackbone=CNN, Fine-tuning Parameters=57M, Evaluation Protocol=fine-tuning2026.0545
CAV-MAE#Param=86M, Pre-training Data=AS+IN, Pre-training Type=Self-Supervised Pre-Training2022.1244.9
PSLA#Param=14M, Pre-training Data=IN, Pre-training Type=Out-of-domain Supervised Pre-Training2022.1244.4
PSLA# Param=14M, Pre-training data=AS2023.0644.4
MBT#Param=86M, Pre-training Data=IN-21K, Pre-training Type=Out-of-domain Supervised Pre-Training2022.1244.3
MBTBackbone=ViT-B/16, Pre-training Data=IN-21K, Pre-training Parameters=343M, Fine-tuning Parameters=86M, Evaluation Protocol=fine-tuning2026.0544.3
PANN# Param=81M, Pre-training data=AS2023.0643.9
ImageBindBackbone=ViT-H2024.0643.4
PANN#Param=81M, Pre-training Type=No Pre-Training2022.1243.1
PANNBackbone=CNN, Fine-tuning Parameters=81M, Evaluation Protocol=fine-tuning2026.0543.1
BAT# Params=91M, Data=AS-2M, Source=Our tuning/probing, Evaluation Protocol=Convex Gated Probing (CGP)2026.0242.9
EAT-S/MFine-tuning=true, # Params=25.5M2024.0442.6
Conformer-Based# Param=88M, Pre-training data=67K hours2023.0641.5
Conformer-SSLBackbone=Conformer, Pre-training Data=AS, Pre-training Parameters=88M, Fine-tuning Parameters=88M, Evaluation Protocol=fine-tuning2026.0541.1
EAT# Params=88M, Data=AS-2M, Source=Our tuning/probing, Evaluation Protocol=Convex Gated Probing (CGP)2026.0241.04
A-RWKV-B/16Training Protocol=from-scratch, Precision=bf162025.0940.91
SSLAM# Params=88M, Data=AS-2M, Source=Our tuning/probing, Evaluation Protocol=Convex Gated Probing (CGP)2026.0240.18
Meta-TransformerBackbone=ViT-L2024.0638.9
AST-B/16Training Protocol=from-scratch, Precision=bf162025.0935.23
BAT# Params=91M, Data=AS-2M, Source=Our tuning/probing, Evaluation Protocol=Linear Probing (LP)2026.0234.27
AuM-B/16Training Protocol=from-scratch2025.0932.43
EAT# Params=88M, Data=AS-2M, Source=Our tuning/probing, Evaluation Protocol=Linear Probing (LP)2026.0230.36
AST-B/16Training Protocol=from-scratch2025.0929.1
SSLAM# Params=88M, Data=AS-2M, Source=Our tuning/probing, Evaluation Protocol=Linear Probing (LP)2026.0228.92
AudioMosaicProbing mode=Linear probing, Encoder=Frozen2026.0528.7
AudioCLIP#Param=93M, Pre-training Data=TI+AS, Pre-training Type=Out-of-domain Supervised Pre-Training2022.1225.9
AudioCLIPBackbone=ESResNeXt, Pre-training Data=TI+AS, Pre-training Parameters=134M, Fine-tuning Parameters=93M, Evaluation Protocol=fine-tuning2026.0525.9
Audio-MAEProbing mode=Linear probing, Encoder=Frozen2026.0520.5
SSLAMProbing mode=Linear probing, Encoder=Frozen2026.0519.5