How to get state_dict of pretrained weight

[magicffourier]

I was trying to inference the yolov5x model and I found your pretrained weight save the whole model, which make torch.load function rely on the relative path of yolov5x.yaml.
And I have tried to load weight like this:
model = Model(r'models/yolov5x.yaml') trained_model = torch.load(opt.weights, map_location=device)['model'].to(device).eval() model.load_state_dict(trained_model.state_dict()) model.to(device).eval()
But I found that the inference outputs of the two models are different

[github-actions]

Hello @magicffourier, thank you for your interest in our work! Please visit our Custom Training Tutorial to get started, and see our Google Colab Notebook, Docker Image, and GCP Quickstart Guide for example environments.

If this is a bug report, please provide screenshots and minimum viable code to reproduce your issue, otherwise we can not help you.

If this is a custom model or data training question, please note that Ultralytics does not provide free personal support. As a leader in vision ML and AI, we do offer professional consulting, from simple expert advice up to delivery of fully customized, end-to-end production solutions for our clients, such as:

• Cloud-based AI surveillance systems operating on hundreds of HD video streams in realtime.
• Edge AI integrated into custom iOS and Android apps for realtime 30 FPS video inference.
• Custom data training, hyperparameter evolution, and model exportation to any destination.

For more information please visit https://www.ultralytics.com.

[glenn-jocher]

@magicffourier yes, checkpoints are saved as full models. The alternative is to save a state_dict, which the user would need to manually pair with a configuration file. We use the 2 file approach in https://github.com/ultralytics/yolov3 but have moved away from this on purpose, as too many people would fail to supply correct pairs of files, then raise issues on the repo for what they perceived as a bug when their incorrect pairing triggered errors (i.e. python detect.py --weights yolov3-spp.pt --cfg yolov3.cfg)

[github-actions]

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.