In PyTorch 2.6, we changed the default value of the weights_only argument in torch.

问题描述：

D:\anaconda\envs\yolov5_mogui\python.exe D:/Instance_Segmentation/yolov5_moguimianju/yolov5-seg-master/train.py
train: weights=weights\yolov5s-seg.pt, cfg=models/yolov5s-seg.yaml, data=data\dc.yaml, hyp=data\hyps\hyp.scratch-low.yaml, epochs=1, batch_size=4, imgsz=640, rect=False, resume=False, nosave=False, noval=False, noautoanchor=False, noplots=False, evolve=None, bucket=, cache=None, image_weights=False, device=0, multi_scale=False, single_cls=False, optimizer=SGD, sync_bn=False, workers=4, project=runs\train-seg, name=exp, exist_ok=False, quad=False, cos_lr=False, label_smoothing=0.0, patience=100, freeze=[0], save_period=-1, seed=0, local_rank=-1, mask_ratio=4, no_overlap=False
github: skipping check (not a git repository), for updates see https://github.com/ultralytics/yolov5
YOLOv5  2025-5-4 Python-3.9.21 torch-2.7.0+cu118 CUDA:0 (NVIDIA GeForce RTX 4060 Laptop GPU, 8188MiB)

hyperparameters: lr0=0.01, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=0.05, cls=0.5, cls_pw=1.0, obj=1.0, obj_pw=1.0, iou_t=0.2, anchor_t=4.0, fl_gamma=0.0, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=0.0, translate=0.1, scale=0.5, shear=0.0, perspective=0.0, flipud=0.0, fliplr=0.5, mosaic=1.0, mixup=0.0, copy_paste=0.0
TensorBoard: Start with 'tensorboard --logdir runs\train-seg', view at http://localhost:6006/
Traceback (most recent call last):
  File "D:\Instance_Segmentation\yolov5_moguimianju\yolov5-seg-master\train.py", line 658, in <module>
    main(opt)
  File "D:\Instance_Segmentation\yolov5_moguimianju\yolov5-seg-master\train.py", line 554, in main
    train(opt.hyp, opt, device, callbacks)
  File "D:\Instance_Segmentation\yolov5_moguimianju\yolov5-seg-master\train.py", line 114, in train
    ckpt = torch.load(weights, map_location='cpu')  # load checkpoint to CPU to avoid CUDA memory leak
  File "D:\anaconda\envs\yolov5_mogui\lib\site-packages\torch\serialization.py", line 1524, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
    (1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
    (2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
    WeightsUnpickler error: Unsupported global: GLOBAL models.yolo.SegmentationModel was not an allowed global by default. Please use `torch.serialization.add_safe_globals([models.yolo.SegmentationModel])` or the `torch.serialization.safe_globals([models.yolo.SegmentationModel])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.

Process finished with exit code 1

问题解答：

降级 PyTorch 版本，可以100%绕过你目前的这个坑。

---

🔧 核心逻辑：

• 你的 yolov5-seg-master 代码其实是兼容 torch 1.13 ~ 2.5 这一段。

• 从 torch 2.6 开始，加入了新的反序列化安全机制（weights_only=True），你没修改代码的话，它在加载模型权重时就会报这个 UnpicklingError。

• 所以降到 2.5.x 就完全回到老逻辑，模型文件可以直接正常读入。

---

注意：

下载torch一定要从官网下载，不要直接pip install torch.