Video Classification on HMDB 51 (test)
76.9AccuracyUniFormer
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| UniFormerPretrain=IN1K+K400, Backbone=UniFormer-S2022.01 | 76.9 | — | — | |
| TDNPretrain=IN1K+K400, Backbone=ResNet502022.01 | 76.3 | — | — | |
| I3DPretrain=IN1K+K400, Backbone=InceptionV22022.01 | 74.8 | — | — | |
| R(2+1)DPretrain=K400, Backbone=ResNet342022.01 | 74.5 | — | — | |
| VidTrPretrain=IN21K+K400, Backbone=ViT-B2022.01 | 74.4 | — | — | |
| TEAPretrain=IN1K+K400, Backbone=ResNet502022.01 | 73.3 | — | — | |
| CT-NetPretrain=IN1K+K400, Backbone=ResNet502022.01 | 73.2 | — | — | |
| STMPretrain=IN1K+K400, Backbone=ResNet502022.01 | 72.2 | — | — | |
| TSMPretrain=IN1K+K400, Backbone=ResNet502022.01 | 70.7 | — | — | |
| C3DPretrain=Sports-1M, Backbone=ResNet182022.01 | 54.9 | — | — | |
| ActionCLIPevaluation_protocol=zero-shot, view=single-view2022.12 | — | 40.8 | — | |
| CoCaevaluation_protocol=zero-shot, view=single-view, aggregation=frame-level average pooling2022.12 | — | 57.2 | 82.6 | |
| ER-ZSARevaluation_protocol=zero-shot, view=single-view2022.12 | — | 35.3 | — | |
| MOV-Levaluation_protocol=zero-shot, view=single-view2022.12 | — | 57.8 | — | |
| RESTevaluation_protocol=zero-shot, view=single-view2022.12 | — | 49.7 | — | |
| VideoCoCaevaluation_protocol=zero-shot, view=single-view, continued_pretraining=true2022.12 | — | 58.7 | 84.5 | |
| VideoCoCa (w.o. videos)evaluation_protocol=zero-shot, view=single-view, continued_pretraining=false2022.12 | — | 57.4 | 82.7 | |
| X-CLIPevaluation_protocol=zero-shot, view=single-view2022.12 | — | 44.6 | — | |
| X-Florenceevaluation_protocol=zero-shot, view=single-view2022.12 | — | 48.4 | — |