Skip to main content
. 2019 Jan 23;2019:7320365. doi: 10.1155/2019/7320365

Figure 3.

Figure 3

The best architecture for MNN with the training method (trainlm).