Missing key(s) in state_dict: "total_ops", "total_params", "module_list.total_ops"

[daixiangzi]

Before submitting a bug report, please be aware that your issue must be reproducible with all of the following, otherwise it is non-actionable, and we can not help you:

• Current repo: run git fetch && git status -uno to check and git pull to update repo
• Common dataset: coco2017.data or coco64.data
• Common environment: Colab, Google Cloud, or Docker image. See https://github.com/ultralytics/yolov3#reproduce-our-environment --Comme If this is a custom dataset/training question you must include* your `train.jpg,test*.jpgandresults.pngfigures, or we can not help you. You can generate these withutils.plot_results()`.

🐛 Bug

A clear and concise description of what the bug is. I find total_ops is registed model,when use thop in file utils/torch_utils.py .so i load weight files ,model exist key total_ops ,but weight not exist,so come up error.

To Reproduce (REQUIRED)

Cmd:

python3 test.py --cfg yolov3-spp.cfg --weights weights/yolov3-spp-ultralytics.pt
Namespace(augment=False, batch_size=16, cfg='./cfg/yolov3-spp.cfg', conf_thres=0.001, data='data/coco2014.data', device='', img_size=512, iou_thres=0.6, save_json=True, single_cls=False, task='test', weights='weights/yolov3-spp-ultralytics.pt')
Model Summary: 225 layers, 6.29987e+07 parameters, 6.29987e+07 gradients, 118.0 GFLOPS

Traceback (most recent call last): File "test.py", line 263, in opt.augment) File "test.py", line 40, in test model.load_state_dict(torch.load(weights, map_location=device)['model']) File "/anaconda3/envs/pytorch-gpu/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1045, in load_state_dict self.class.name, "\n\t".join(error_msgs))) RuntimeError: Error(s) in loading state_dict for Darknet: Missing key(s) in state_dict: "total_ops", "total_params", "module_list.total_ops", "module_list.total_params", "module_list.0.total_ops"

## Expected behavior
A clear and concise description of what you expected to happen.

## Environment
If applicable, add screenshots to help explain your problem.

 - OS: Ubuntu20
 - GPU 2080 Ti
python3.7

## Additional context
Add any other context about the problem here.

[daixiangzi]

solved method: 1 model.load_state_dict(state_dice,strick=False) 2 pip3 uninstall thop or conda uninstall thop

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[ZhangYuef]

Another approach for this issue: to delete unmatched total_ops and total_params in state_dict's keys

    checkpoint = torch.load(cpt_path)
    # delete unmatched total_ops total_params
    state_dict = []
    for n, p in checkpoint['model'].items():
        if "total_ops" not in n and "total_params" not in n:
            state_dict.append((n, p))
    state_dict = dict(state_dict)

    model.load_state_dict(state_dict)

[Jingnan-Jia]

solved method: 1 model.load_state_dict(state_dice,strick=False) 2 pip3 uninstall thop or conda uninstall thop

a typo: strick=False should be strict=False

[pvti]

In my case, it related to thop.

[pvti]

@ZhangYuef your solution woked in my case