Action Recognition on Something-Something V1 (test val)
57.2Top-1 AccTCM-R50 En
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TCM-R50 EnFrame=16+8, FLOPs x clips=105G x 10, Params=49.0M2022.02 | 57.2 | 85.2 | |
| TDNFrame=16+8, FLOPs x clips=198G x 12022.02 | 56.8 | 84.1 | |
| MSNetFrame=16+8, FLOPs x clips=101G x 10, Params=49.2M2022.02 | 55.1 | 84 | |
| TCM-R50 EnFrame=16+8, FLOPs x clips=105G x 1, Params=49.0M2022.02 | 54.7 | 82.6 | |
| CAN#Frame=16, Params (M)=25.3, FLOPs (G)=70, Clips=12025.12 | 53.9 | 82 | |
| TCM-R50Frame=16, FLOPs x clips=70G x 1, Params=24.5M2022.02 | 53.1 | 81.2 | |
| TEINetFrame=16+8, FLOPs x clips=99G x 1, Params=30.4M2022.02 | 52.5 | — | |
| TCM-R50Frame=8, FLOPs x clips=35G x 1, Params=24.5M2022.02 | 52.2 | 80.4 | |
| TEAFrame=16, FLOPs x clips=70G x 1, Params=24.5M2022.02 | 51.9 | 80.3 | |
| TEA#Frame=16, Params (M)=24.5, FLOPs (G)=70, Clips=12025.12 | 51.9 | 80.3 | |
| CorrNet-101Frame=32, FLOPs x clips=224G x 302022.02 | 51.7 | — | |
| CorrNet-101#Frame=32, FLOPs (G)=224, Clips=302025.12 | 51.7 | — | |
| ESTI#Frame=16, FLOPs (G)=66, Clips=12025.12 | 51.7 | 81.3 | |
| GC-TSN#Frame=16, Params (M)=25.1, FLOPs (G)=67, Clips=12025.12 | 51.3 | 80 | |
| TDRL#Frame=16, FLOPs (G)=66, Clips=12025.12 | 50.9 | — | |
| STMFrame=16, FLOPs x clips=67G x 30, Params=24.0M2022.02 | 50.7 | 80.4 | |
| TANetFrame=16+8, FLOPs x clips=99G x 1, Params=25.6M2022.02 | 50.6 | 79.3 | |
| CAN#Frame=8, Params (M)=25.3, FLOPs (G)=35, Clips=12025.12 | 50.4 | 78.2 | |
| 3D DenseNet121Frame=16, FLOPs x clips=31G x 1, Params=21.4M2022.02 | 50.2 | 78.9 | |
| ESTI#Frame=8, FLOPs (G)=33, Clips=12025.12 | 50.2 | 79.8 | |
| DFB-NetFrame=16, FLOPs x clips=N/A x 12022.02 | 50.1 | 79.5 | |
| GENet#Frame=16, Params (M)=24.5, FLOPs (G)=67, Clips=12025.12 | 50.1 | 79 | |
| TDRL#Frame=8, FLOPs (G)=33, Clips=12025.12 | 49.8 | — | |
| TSMFrame=16+8, FLOPs x clips=98G x 1, Params=48.6M2022.02 | 49.7 | 78.5 | |
| GC-TSN#Frame=8, Params (M)=25.1, FLOPs (G)=33, Clips=12025.12 | 49.7 | 78.2 | |
| SmallBigNet#Frame=16, FLOPs (G)=105, Clips=12025.12 | 49.3 | 79.5 | |
| TEA#Frame=8, Params (M)=24.5, FLOPs (G)=35, Clips=12025.12 | 48.9 | 78.1 | |
| GENet#Frame=8, Params (M)=24.5, FLOPs (G)=33, Clips=12025.12 | 48.7 | 77.5 | |
| S3D-GFrame=64, FLOPs x clips=71G x 1, Params=11.6M2022.02 | 48.2 | 78.7 | |
| S3D-G#Frame=64, Params (M)=11.6, FLOPs (G)=71, Clips=22025.12 | 48.2 | 78.7 | |
| TANet#Frame=16, Params (M)=25.1, FLOPs (G)=66, Clips=12025.12 | 47.6 | 77.7 | |
| TSM#Frame=16, Params (M)=24.3, FLOPs (G)=65, Clips=12025.12 | 47.3 | 77.1 | |
| SmallBigNet#Frame=8, FLOPs (G)=52, Clips=12025.12 | 47 | 77.1 | |
| TANet#Frame=8, Params (M)=25.1, FLOPs (G)=33, Clips=12025.12 | 46.5 | 75.8 | |
| ECO En LiteFrame=92, FLOPs x clips=267 x 1, Params=150M2022.02 | 46.4 | — | |
| RubiksNetFrame=8, FLOPs x clips=33G x 12022.02 | 46.4 | 74.5 | |
| RubiksNet#Frame=8, FLOPs (G)=33, Clips=12025.12 | 46.4 | 74.5 | |
| NL-I3D + GCNFrame=32, FLOPs x clips=303G x 2, Params=62.2M2022.02 | 46.1 | 76.8 | |
| NL-I3D+GCN#Frame=32, Params (M)=62.2, FLOPs (G)=303, Clips=22025.12 | 46.1 | 76.8 | |
| TSM#Frame=8, Params (M)=24.3, FLOPs (G)=33, Clips=12025.12 | 45.6 | 74.2 | |
| NL-I3DFrame=32, FLOPs x clips=168G x 2, Params=35.3M2022.02 | 44.4 | 76 | |
| NL-I3D#Frame=32, Params (M)=35.3, FLOPs (G)=168, Clips=22025.12 | 44.4 | 76 | |
| MFNetFrame=10, FLOPs x clips=N/A x 102022.02 | 43.9 | 73.1 | |
| I3DFrame=32, FLOPs x clips=153G x 2, Params=28.0M2022.02 | 41.6 | 72.2 | |
| I3D#Frame=32, Params (M)=28.0, FLOPs (G)=153, Clips=22025.12 | 41.6 | 72.2 | |
| TRNFrame=8, FLOPs x clips=16G x N/A, Params=18.3M2022.02 | 34.4 | — | |
| TSN#Frame=8, Params (M)=23.9, FLOPs (G)=33, Clips=12025.12 | 19.7 | — | |
| TSNFrame=8, FLOPs x clips=16G x 1, Params=10.7M2022.02 | 19.5 | — |