We utilized this to avoid overfitting as we trained only with the larger CMP ... used the Scaled YOLOv4 network [31], which we already mentioned in Sect. 1.
確定! 回上一頁