= [10, 13, 16, 30, 33, 23, 30, 61, 62, 45, 59, 119, 116, 90, 156, 198, 373, 326], anchor_masks...downsample = 32 for i, out in enumerate(outputs): # 对三个层级分别求损失函数 anchor_mask_i = anchor_masks...as fluid ANCHORS = [10, 13, 16, 30, 33, 23, 30, 61, 62, 45, 59, 119, 116, 90, 156, 198, 373, 326] ANCHOR_MASKS...=ANCHOR_MASKS, valid_thresh = VALID_THRESH) bboxes_data...=ANCHOR_MASKS, valid_thresh = VALID_THRESH) bboxes_data
conv_group_scale: 1 with_extra_blocks: false YOLOv3Head: anchor_masks: [[6, 7, 8], [3, 4, 5], [0
conv_group_scale: 1 with_extra_blocks: false YOLOv3Head: anchor_masks: [[6, 7, 8], [3, 4, 5], [0,
anchor_masks (list|tuple): 在计算YOLOv3损失时,使用anchor的mask索引,为None时表示使用默认值 [[6, 7, 8], [3, 4, 5], [0, 1, 2
_darknet_conv(x, 512 * 2, 3) # 9个簇,3个尺度 anchor_masks = np.array([[6, 7, 8], [3, 4, 5], [0, 1, 2]]) #...3 num_anchors = len(anchor_masks[0]) # [batch, h, w, num_anchors * (num_class + 5)] # 此处不使用批归一化和激活函数,
,真实框所属类别,维度是[N, 50] iou_threshold,当预测框与真实框的iou大于iou_threshold时不将其看作是负样本 anchors,锚框可选的尺寸 anchor_masks
领取专属 10元无门槛券
手把手带您无忧上云