📊 Showing 5 results | 📏 Metric: 3-fold Accuracy
| Rank | Model | Paper | 3-fold Accuracy | Date | Code |
|---|---|---|---|---|---|
| 1 | STM (ImageNet+Kinetics pretrain) | STM: SpatioTemporal and Motion Encoding for Action Recognition | 96.20 | 2019-08-07 | - |
| 2 | 3D-SqueezeNet | Resource Efficient 3D Convolutional Neural Networks | 74.94 | 2019-04-04 | 📦 okankop/Efficient-3DCNNs 📦 ahmetgunduz/Real-time-GesRec |
| 3 | 3D-ShuffleNetV2 0.25x | Resource Efficient 3D Convolutional Neural Networks | 56.52 | 2019-04-04 | 📦 okankop/Efficient-3DCNNs 📦 ahmetgunduz/Real-time-GesRec |
| 4 | 3D-MobileNetV2 0.2x | Resource Efficient 3D Convolutional Neural Networks | 55.56 | 2019-04-04 | 📦 okankop/Efficient-3DCNNs 📦 ahmetgunduz/Real-time-GesRec |
| 5 | Baseline UCF101 | UCF101: A Dataset of 101 Human Actions Classes From Videos in The Wild | 43.90 | 2012-12-03 | 📦 wushidonguc/two-stream-action-recognition-keras 📦 doronharitan/human_activity_recognition_LRCN 📦 ryanchesler/comma-speed-challenge |