Research Article
[Retracted] Bio-Optimization of Deep Learning Network Architectures
Table 2
The performance of DNN architecture using Fashion MNIST dataset.
| DNN architecture | Dataset | Epochs | Batch size | Optimizers | Training accuracy | Testing accuracy | Loss |
| Fashion MNIST dataset | 5 | 32 | SGD | 88.43 | 76.34 | 0.32 | RMSprop | 87.34 | 77.32 | 0.41 | Adam | 84.69 | 78.63 | 0.48 | Adadelta | 83.54 | 74.64 | 0.30 | Adagrad | 85.28 | 75.85 | 0.34 | Adamax | 86.37 | 74.73 | 0.38 | NAdam | 86.35 | 76.84 | 0.37 | Ftrl | 89.55 | 79.68 | 0.52 |
|
|