Human Activity Recognition on UT-HAR
99.2AccuracyLightweight TCN
Evaluation results
| Method | Links | |
|---|---|---|
| Lightweight TCNFLOPS (M)=30.68, Parameters (M)=0.2522026.06 | 99.2 | |
| Liu et al.FLOPS (M)=89.87, Parameters (M)=12.6102026.06 | 99.13 | |
| Tan et al.FLOPS (M)=142.32, Parameters (M)=0.6192026.06 | 99.02 | |
| Luo et al.Parameters (M)=1.1602026.06 | 98.78 | |
| Li et al.FLOPS (M)=70.97, Parameters (M)=1.7602026.06 | 98.44 | |
| Zhang et al.Parameters (M)=0.422026.06 | 98.19 | |
| ResNet18*FLOPS (M)=49.93, Parameters (M)=11.1802026.06 | 98.11 | |
| ResNet50*FLOPS (M)=86.40, Parameters (M)=23.5502026.06 | 97.21 | |
| ViT*FLOPS (M)=273.10, Parameters (M)=10.5802026.06 | 96.53 | |
| GRU*FLOPS (M)=7.60, Parameters (M)=0.0302026.06 | 94.18 | |
| BiLSTM*FLOPS (M)=20.29, Parameters (M)=0.0802026.06 | 90.19 | |
| LSTM*FLOPS (M)=10.14, Parameters (M)=0.0402026.06 | 87.18 | |
| RNN*FLOPS (M)=2.51, Parameters (M)=0.0102026.06 | 83.53 |