Table 4 Comparison between deferent number of stacked self-attention blocks.
From: An efficient self-attention network for skeleton-based action recognition
N | cs (%) | cv (%) |
|---|---|---|
2 | 73.0 | 80.0 |
3 | 86.5 | 90.3 |
4 | 90.6 | 96.1 |
5 | 90.5 | 96.1 |
From: An efficient self-attention network for skeleton-based action recognition
N | cs (%) | cv (%) |
|---|---|---|
2 | 73.0 | 80.0 |
3 | 86.5 | 90.3 |
4 | 90.6 | 96.1 |
5 | 90.5 | 96.1 |
