Table 1 Neural network configurations. Designations in parentheses correspond to those in Fig. 3A
From: Marigold: a machine learning-based web app for zebrafish pose tracking
Dataset | Touch-evoked | Visuomotor |
|---|---|---|
Input resolution (H × W) | 512 × 512 | 256 × 256 |
Output resolution (H/2 × W/2) | 256 × 256 | 128 × 128 |
Input channels (I) | 1 | 1 |
Middle channels (M) | 32 | 16 |
Output channels (O) | 7 | 1 |
Residual blocks (N) | 10 | 10 |