Table 6 Comparison table of various model parameters.

From: TCN-attention-HAR: human activity recognition based on attention mechanism time convolutional network

Method

Model parameters

Training time per epoch

GRU-attention

2,163,340

18.797

CNN-GRU

267,404

6.974

Attention-LSTM

2,180,240

18.832

BiLSTM

70.668

4.212

CNN-LSTM

348,810

7.048

CNN-BiLSTM

874,124

7.160

CNN-attention-BiLSTM

4,605,068

14.940

CNN-BiGRU

662,156

7.717

TAHAR-student-CNN

17,804

3066

TAHAR-student-LSTM

18,956

3.480

TAHAR-student-GRU

14,604

3.472

TAHAR-teacher

10,950,162

36.319

  1. Significant values are in bold.
Back to article page