war0927

In the parameter settings, the opt.batch_size is set to 8, but why the shape[0] of the feature map is 1 during training, do you have any friends who can explain...

关于包

3

from detrex.layers import MLP, box_cxcywh_to_xyxy, box_xyxy_to_cxcywh from detrex.utils import inverse_sigmoid from detectron2.modeling import detector_postprocess from detectron2.structures import Boxes, ImageList, Instances 其中detectron2、inverse_sigmoid、MLP, box_cxcywh_to_xyxy, box_xyxy_to_cxcywh显示没有

ValueError: Expected more than 1 value per channel when training, got input size torch.Size([1, 256, 1, 1])

2

我采用你的方法进行特征融合， nn.AdaptiveAvgPool2d(1), nn.Conv2d(channels, inter_channels, kernel_size=1, stride=1, padding=0), nn.BatchNorm2d(inter_channels), nn.ReLU(inplace=True), nn.Conv2d(inter_channels, channels, kernel_size=1, stride=1, padding=0), nn.BatchNorm2d(channels), 第一行将自适应池化的输出按照原文设为1时会报错ValueError: Expected more than 1 value per channel when training, got input size torch.Size([1, 256,...

关于BN层

2

请问当我应用iaff模块时候存在以下情况：1、F的batch_size为1；2、全局平均池化将特征图的面积变为1*1。这导致了经过BN层的时候会报错ValueError: Expected more than 1 value per channel when training, got input size torch.Size([1, 256, 1, 1])，请问这个有处理方法吗