JohnFiv

同问

> 看文中对coco数据集的small，medium，large目标的标注个数及其在图像上的分布，想知道怎么做到的 请问你找到了吗

> 你说我刚改的这个么？ 不是刚改的，刚改的我还没跑，是之前2.0版本的，loss换成ciou会变差很多

loss_x = torch.mean(self.BCELoss(x[obj_mask], y_true[..., 0][obj_mask]) * box_loss_scale) RuntimeError: The size of tensor a (5) must match the size of tensor b (13) at non-singleton dimension 3

Traceback (most recent call last): File "D:\Pytorch_workspace\yolo3-pytorch-master_1\utils\utils_fit.py", line 42, in fit_one_epoch loss_item = yolo_loss(l, outputs[l], targets) File "D:\ProgramData\Anaconda3\envs\pytorch\lib\site-packages\torch\nn\modules\module.py", line 547, in __call__ result = self.forward(*input, **kwargs) File "D:\Pytorch_workspace\yolo3-pytorch-master_1\nets\yolo_training.py", line 192,...