S3D-G (ImageNet, Kinetics-400 pretrained)

Reported on 4 benchmarks across 2 tasks · 1 paper

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Robots2 results

Activity RecognitiononHMDB-51
Average accuracy of 3 splits· 2017-12-13
75.9
best: 88.7 (VideoMAE V2-g)
Rethinking Spatiotemporal Feature Learning: Speed-Accuracy Trade-offs in Video Classification arXiv:1712.04851
Activity RecognitiononUCF101
3-fold Accuracy· uses extra data· 2017-12-13
96.8
best: 99.7 (FTP-UniFormerV2-L/14)
Rethinking Spatiotemporal Feature Learning: Speed-Accuracy Trade-offs in Video Classification arXiv:1712.04851

Time Series2 results

Action RecognitiononHMDB-51
Average accuracy of 3 splits· 2017-12-13
75.9
best: 88.7 (VideoMAE V2-g)
Rethinking Spatiotemporal Feature Learning: Speed-Accuracy Trade-offs in Video Classification arXiv:1712.04851
Action RecognitiononUCF101
3-fold Accuracy· uses extra data· 2017-12-13
96.8
best: 99.7 (FTP-UniFormerV2-L/14)
Rethinking Spatiotemporal Feature Learning: Speed-Accuracy Trade-offs in Video Classification arXiv:1712.04851