在使用YOLO框架训练自己的数据集时候,开始跑train.py,出现如下报错:
RuntimeError: result type Float can‘t be cast to the desired output type long int
Traceback (most recent call last): File "/home/sjh/project/yolov5-5.0/train.py", line 543, in <module> train(hyp, opt, device, tb_writer) File "/home/sjh/project/yolov5-5.0/train.py", line 304, in train loss, loss_items = compute_loss(pred, targets.to(device)) # loss scaled by batch_size File "/home/sjh/project/yolov5-5.0/utils/loss.py", line 117, in __call__ tcls, tbox, indices, anchors = self.build_targets(p, targets) # targets File "/home/sjh/project/yolov5-5.0/utils/loss.py", line 211, in build_targets indices.append((b, a, gj.clamp_(0, gain[3] - 1), gi.clamp_(0, gain[2] - 1))) # image, anchor, grid indices RuntimeError: result type Float can't be cast to the desired output type long int
由于是初次使用,并不确定是哪里的问题,只知道是跟着其他人的教程来的,经过查询,得到问题原因:
官网的yolov5-master可以正常运行,但是yolov5-5.0/yolov5-6.1等版本会出问题;
这是因为yolov5-master版本和yolov5-5.0/yolov5-6.1等版本下的【utils】中的【loss.py】文件不同,大概是yolov5-5.0/yolov5-6.1等版本在更新版本的时候出了问题
因此解决方法也明了了:修改loss.py文件
打开【utils】文件夹下的【loss.py】文件,一共需要修改两处代码,以yolov5-5.0为例:
1,大概是177行左右,可通过搜索找到下面这句代码
anchors = self.anchors[i]
将这一行其替换为
anchors, shape = self.anchors[i], p[i].shape
2,在程序的最后,找到下面这行代码
indices.append((b, a, gj.clamp_(0, gain[3] - 1), gi.clamp_(0, gain[2] - 1))) # image, anchor, grid indices
将其替换为
indices.append((b, a, gj.clamp_(0, shape[2] - 1), gi.clamp_(0, shape[3] - 1))) # image, anchor, grid
保存程序,重新跑train.py,便可以正常运行了