Table 7 Improved parameter setting of DCGAN sample augmentation model.
Parameter | Numerical value |
|---|---|
Learning rate | 0.002 |
Optimization strategy | Adam |
Batch size | 128 |
Loss function | Cross entropy |
Number of pre-generated samples of each type | 5000 |
Parameter | Numerical value |
|---|---|
Learning rate | 0.002 |
Optimization strategy | Adam |
Batch size | 128 |
Loss function | Cross entropy |
Number of pre-generated samples of each type | 5000 |