[Retracted] Gradient Descent Optimization in Deep Learning Model Training Based on Multistage and Method Combination Strategy

<div>Comparison between a conflict operation and a single operation of learning rate on CIFAR10. (a, b) The comparison result of LSTM, and (c, d) the comparison result of ResNet-20.</div>

Security and Communication Networks

fig5

Figure 5

Figure 5: [Retracted] Gradient Descent Optimization in Deep Learning Model Training Based on Multistage and Method Combination Strategy 