Video Classification on COIN (test)
94.1Top-1 AccuracyFluxViT
Evaluation Results
| Method | Links | |
|---|---|---|
| FluxViTe2e=true, Backbone=FluxViT-B, Token Number=3072, Token Selection Strategy=More informative tokens (Token Optimization)2025.03 | 94.1 | |
| FluxViTe2e=true, Backbone=FluxViT-B, Token Number=3072, Token Selection Strategy=Standard (unmasked frames)2025.03 | 93.9 | |
| FluxViTe2e=true, Backbone=FluxViT-B, Token Number=2048, Token Selection Strategy=More informative tokens (Token Optimization)2025.03 | 93.9 | |
| FluxViTe2e=true, Backbone=FluxViT-B, Token Number=2048, Token Selection Strategy=Standard (unmasked frames)2025.03 | 93.7 | |
| HERMESe2e=true, Backbone=MLLM2025.03 | 93.5 | |
| MA-LMMe2e=true, Backbone=MLLM2025.03 | 93.2 | |
| FluxViTe2e=true, Backbone=FluxViT-B, Token Number=1024, Token Selection Strategy=More informative tokens (Token Optimization)2025.03 | 93.2 | |
| FluxViTe2e=true, Backbone=FluxViT-B, Token Number=1024, Token Selection Strategy=Standard (unmasked frames)2025.03 | 92.5 | |
| FluxViTe2e=true, Backbone=FluxViT-S, Token Number=3072, Token Selection Strategy=More informative tokens (Token Optimization)2025.03 | 92.1 | |
| FluxViTe2e=true, Backbone=FluxViT-S, Token Number=2048, Token Selection Strategy=More informative tokens (Token Optimization)2025.03 | 91.9 | |
| FluxViTe2e=true, Backbone=FluxViT-S, Token Number=3072, Token Selection Strategy=Standard (unmasked frames)2025.03 | 91.8 | |
| FluxViTe2e=true, Backbone=FluxViT-S, Token Number=2048, Token Selection Strategy=Standard (unmasked frames)2025.03 | 91.5 | |
| FluxViTe2e=true, Backbone=FluxViT-S, Token Number=1024, Token Selection Strategy=More informative tokens (Token Optimization)2025.03 | 91 | |
| VideoMamba_f64e2e=true, Backbone=VideoMamba-M2025.03 | 90.4 | |
| Distant Supervisione2e=false, Backbone=TimeSformer2025.03 | 90 | |
| InternVideo2_f12e2e=true, Backbone=InternVideo2-S2025.03 | 90 | |
| FluxViTe2e=true, Backbone=FluxViT-S, Token Number=1024, Token Selection Strategy=Standard (unmasked frames)2025.03 | 89.8 | |
| VideoMamba_f64e2e=true, Backbone=VideoMamba-S2025.03 | 88.7 | |
| ViS4mere2e=false, Backbone=Swin-B2025.03 | 88.4 | |
| Turbo_f32e2e=true, Backbone=VideoMAE-B2025.03 | 87.5 |