yolo系列全部有问题，训练解码结果不对

[williamlzw]

yolo系列全部有问题，不加载预训练模型，解码结果偏差很大

[bubbliiiing]

……你训练了吗？数据集很大吗？

[williamlzw]

训练集40张，测试集10张，2分类。loss降到3.多，结果全错。yolo3,yolo4也训练不出来。我用别的项目模型训练矩形框偏差不大

[williamlzw]

有qq群，微信群吗

[bubbliiiing]

……你看注释了？

[williamlzw]

Freeze_Train = False，不进行冻结训练，我移植模型和解码到torchsharp，训练效果也不好。

[bubbliiiing]

……说了啊，没有预训练权重需要注意什么

[Milkyway-xX]

你确定就这几十张图能训练？

[bubbliiiing]

有预训练可以勉强用用

[dstawj]

D:\anaconda3\envs\py1\python.exe F:\core_code\yolov8-pytorch-master\get_map.py Load model. initialize network with normal type Traceback (most recent call last): File "F:\core_code\yolov8-pytorch-master\get_map.py", line 92, in yolo = YOLO(confidence = confidence, nms_iou = nms_iou) File "F:\core_code\yolov8-pytorch-master\yolo.py", line 91, in init self.generate() File "F:\core_code\yolov8-pytorch-master\yolo.py", line 105, in generate self.net.load_state_dict(torch.load(self.model_path, map_location=device)) File "D:\anaconda3\envs\py1\lib\site-packages\torch\nn\modules\module.py", line 1052, in load_state_dict self.class.name, "\n\t".join(error_msgs))) RuntimeError: Error(s) in loading state_dict for YoloBody: size mismatch for cv3.0.2.weight: copying a param with shape torch.Size([80, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 128, 1, 1]). size mismatch for cv3.0.2.bias: copying a param with shape torch.Size([80]) from checkpoint, the shape in current model is torch.Size([1]). size mismatch for cv3.1.2.weight: copying a param with shape torch.Size([80, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 128, 1, 1]). size mismatch for cv3.1.2.bias: copying a param with shape torch.Size([80]) from checkpoint, the shape in current model is torch.Size([1]). size mismatch for cv3.2.2.weight: copying a param with shape torch.Size([80, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([1, 128, 1, 1]). size mismatch for cv3.2.2.bias: copying a param with shape torch.Size([80]) from checkpoint, the shape in current model is torch.Size([1]).

进程已结束，退出代码为 1