You can skip the rest of this tutorial and start training your Mask RCNN model ... HuberLoss() # == smoothl1 # the loss to penalize incorrect segmentation ...
確定! 回上一頁