[GhostNetV3] Question about Figure 3

[Journey7331]

Thanks for your great work!!

The title of Figure 3, "The top-1 validation accuracy for various learning rates of GhostNetV2," appears to be inconsistent with the content, which seems to represent values from Table 7 regarding "weight decay settings."

Additionally, there is a discrepancy in that the text states, "Therefore, a learning rate of 5e-3 is chosen for the final experiments," yet this learning rate is not depicted in Figure 3.

[liu-zhenhua]

Thanks for the attention and the kind reminder. We are sorry for the mistake in Figure 3 and have corrected in the manuscript.