fix bug for trt inference

[AllenZYJ]

When I was doing inference with ”engine“ on “Nvidia agx” device, I found that ”segment/predict.py“ would report a dimension mismatch error. The reason was that the output branch selection was wrong. This simple change fixes that error, and allows smooth inference on trt devices.

🤖 Generated by Copilot at 19c8663

Summary

🐛🚀🧠

Fix output order bug in models/common.py for TensorRT. Hardcode output names to match YOLOv5 output layers.

y variable
hardcoded for TensorRT
autumn bug fixing

Walkthrough

• Fix incorrect detection results when using TensorRT by hardcoding output names to match output layers in YOLOv5 model (link)

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

📊 Key Changes

• Adjusted tensor output retrieval in forward method during TensorRT inference.

🎯 Purpose & Impact

• 🛠️ Fixes a bug in TensorRT inference that could lead to incorrect tensor outputs, ensuring the model's predictions are accurate.
• 🏃‍♂️ Increases reliability for users deploying YOLOv5 models with TensorRT, which can improve performance on compatible hardware.

🌟 Summary

Fixed a crucial bug in YOLOv5's TensorRT inference to ensure correct output tensors are retrieved, bolstering performance and reliability. 🐛➡️🔧

[github-actions]

👋 Hello @AllenZYJ, thank you for submitting a YOLOv5 🚀 PR! To allow your work to be integrated as seamlessly as possible, we advise you to:

• ✅ Verify your PR is up-to-date with ultralytics/yolov5 master branch. If your PR is behind you can update your code by clicking the 'Update branch' button or by running git pull and git merge master locally.
• ✅ Verify all YOLOv5 Continuous Integration (CI) checks are passing.
• ✅ Reduce changes to the absolute minimum required for your bug fix or feature addition. "It is not daily increase but daily decrease, hack away the unessential. The closer to the source, the less wastage there is." — Bruce Lee

[glenn-jocher]

@AllenZYJ thanks for bringing this to our attention! We'll review the proposed changes and assess their compatibility with our general code structure.

It's great to see the community actively working to improve the YOLOv5 codebase for various inference and deployment scenarios. We appreciate your effort and contribution to making YOLOv5 more versatile.

[AllenZYJ]

@AllenZYJ thanks for bringing this to our attention! We'll review the proposed changes and assess their compatibility with our general code structure.

It's great to see the community actively working to improve the YOLOv5 codebase for various inference and deployment scenarios. We appreciate your effort and contribution to making YOLOv5 more versatile.

thanks, it just a simple modified,and I tested it locally.

[glenn-jocher]

@AllenZYJ your contribution is appreciated, AllenZYJ. We value your testing and input, which helps ensure the stability and functionality of YOLOv5 across diverse deployment scenarios.

[AllenZYJ]

@AllenZYJ your contribution is appreciated, AllenZYJ. We value your testing and input, which helps ensure the stability and functionality of YOLOv5 across diverse deployment scenarios.

hello,@glenn-jocher,I was wondering if you've been able to complete testing this PR and if everything passed?

[glenn-jocher]

@AllenZYJ Thanks for your patience. I have not been able to test the PR yet, but I will definitely do so at the earliest opportunity. Your effort and the community's support are invaluable in maintaining the stability and functionality of YOLOv5 in diverse deployment scenarios.

[github-actions]

All Contributors have signed the CLA. ✅
_{Posted by the CLA Assistant Lite bot.}

[AllenZYJ]

I have read the CLA Document and I sign the CLA

[AllenZYJ]

@glenn-jocher please review this merge,thank you very much.

[glenn-jocher]

@AllenZYJ Thank you for your contribution! We'll review the merge as soon as possible. Your patience and efforts are much appreciated! 🙌