Sentiment Analysis on CMU-MOSEI

0.7766WF1WavLM

Evaluation Results

Method
WavLM2023.12	0.7766	—
emotion2vec2023.12	0.7656	—
Whisper Encoder2023.12	0.7646	—
Whisper Encoder2023.12	0.7475	—
WavLM2023.12	0.7347	—
data2vec2023.12	0.7279	—
ConfSMoE-E2025.05	0.6235	0.8275
ConfSMoE-T2025.05	0.6131	0.8287
FlexMoE2025.05	0.6062	0.8134
ConfSMoE-E2025.05	0.6022	0.8105
ConfSMoE-T2025.05	0.5981	0.81
LlMoE2025.05	0.5913	0.8029
FuseMoE-L2025.05	0.5871	0.7813
FlexMoE2025.05	0.5866	0.7948
ConfSMoE-T2025.05	0.5806	0.7885
ConfSMoE-E2025.05	0.5766	0.7883
LlMoE2025.05	0.5746	0.7938
FuseMoE-S2025.05	0.5704	0.7854
mmFormer2025.05	0.5675	0.7642
FuseMoE-L2025.05	0.5668	0.7644
FlexMoE2025.05	0.5596	0.7805
FuseMoE-S2025.05	0.559	0.7638
mmFormer2025.05	0.5537	0.7536
LlMoE2025.05	0.5535	0.7729
ConfSMoE-E2025.05	0.5483	0.771
FuseMoE-L2025.05	0.546	0.7492
FuseMoE-S2025.05	0.5444	0.7432
ConfSMoE-T2025.05	0.5416	0.7691
LlMoE2025.05	0.5393	0.765
mmFormer2025.05	0.5291	0.7302
FuseMoE-S2025.05	0.5227	0.7222
FlexMoE2025.05	0.52	0.7614
LlMoE2025.05	0.5095	0.7324
ConfSMoE-T2025.05	0.5072	0.7363
ConfSMoE-E2025.05	0.5066	0.7388
FuseMoE-L2025.05	0.4992	0.7252
mmFormer2025.05	0.4971	0.7128
FuseMoE-L2025.05	0.4936	0.7005
FuseMoE-S2025.05	0.4932	0.6985
FlexMoE2025.05	0.4929	0.7208
ConfSMoE-E2025.05	0.4813	0.7048
mmFormer2025.05	0.4806	0.6931
ShaSpec2025.05	0.4748	0.6793
ShaSpec2025.05	0.4741	0.6926
mmFormer2025.05	0.4739	0.6661
ConfSMoE-T2025.05	0.4731	0.7174
LlMoE2025.05	0.4669	0.709
ShaSpec2025.05	0.4654	0.6805
FlexMoE2025.05	0.4619	0.7015
ShaSpec2025.05	0.4505	0.6689
FuseMoE-L2025.05	0.4444	0.6744
ShaSpec2025.05	0.4349	0.6532
FuseMoE-S2025.05	0.4308	0.6763
ShaSpec2025.05	0.4275	0.6232
TF2025.05	0.3889	0.6426
TF2025.05	0.3537	0.5954
TF2025.05	0.3536	0.6023
TF2025.05	0.3494	0.6105
TF2025.05	0.3301	0.582
TF2025.05	0.3223	0.6009