R[2+1]D-Flow (Kinetics 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· uses extra data· 2017-11-30
76.4
best: 88.7 (VideoMAE V2-g)
A Closer Look at Spatiotemporal Convolutions for Action Recognition arXiv:1711.11248
Activity RecognitiononUCF101
3-fold Accuracy· uses extra data· 2017-11-30
95.5
best: 99.7 (FTP-UniFormerV2-L/14)
A Closer Look at Spatiotemporal Convolutions for Action Recognition arXiv:1711.11248

Time Series2 results

Action RecognitiononHMDB-51
Average accuracy of 3 splits· uses extra data· 2017-11-30
76.4
best: 88.7 (VideoMAE V2-g)
A Closer Look at Spatiotemporal Convolutions for Action Recognition arXiv:1711.11248
Action RecognitiononUCF101
3-fold Accuracy· uses extra data· 2017-11-30
95.5
best: 99.7 (FTP-UniFormerV2-L/14)
A Closer Look at Spatiotemporal Convolutions for Action Recognition arXiv:1711.11248