Research Article
Detecting Gear Surface Defects Using Background-Weakening Method and Convolutional Neural Network
Table 2
Training part of the main parameter settings.
| Parameter | Numerical value | Parameter | Numerical value | Parameter | Numerical value | Parameter | Numerical value |
| Batch | 64 | Angle | 0 | Burn_in | 2000 | Scales | 0.1, 0.1 | Subdivisions | 32 | Saturation | 1.5 | Max_batches | 50000 | Learning_rate | 0.001 | Momentum | 0.9 | Exposure | 1.5 | Policy | Steps | Random | 1 | Decay | 0.0008 | Hue | 0.1 | Steps | 4000, 4500 | Jitter | 3 |
|
|