= fluid.layers.sigmoid_cross_entropy_with_logits(pred_classification, label_classification)
在前面几个小节中我们已经知道怎么计算这些预测值和标签了...iou_above_thresh_indices
label_objectness[ignore_indices] = -1
return label_objectness
下面通过调用这两个函数,实现如何将部分预测框的...objectness标签设置为-1了,不计算其对任何一种损失函数的贡献。...tw = reshaped_output[:, :, 2, :, :]
th = reshaped_output[:, :, 3, :, :]
# 从label_location中取出各个位置坐标的标签..., [label, score, x1, x2, y1, y2]]
预测框列表中每个元素[label, score, x1, x2, y1, y2]描述了一个预测框,label是预测框所属类别标签,score