Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

torch.split() replace slicing on out-of-place inference #7258

Merged
merged 1 commit into from
Apr 3, 2022

Conversation

glenn-jocher
Copy link
Member

@glenn-jocher glenn-jocher commented Apr 3, 2022

Simplified export, similar profiling on PyTorch inference.

🛠️ PR Summary

Made with ❤️ by Ultralytics Actions

🌟 Summary

Improved tensor operations in YOLOv5 model code.

📊 Key Changes

  • Replaced a set of tensor operations with a more concise and readable split operation (y.split method).
  • The bounding box predictions (xy and wh) and confidence score calculations are now more clearly separated by using the split operation.
  • The updated code now concatenates the separated tensors back together after processing.

🎯 Purpose & Impact

  • The purpose is to enhance code readability and potentially efficiency by using the more modern tensor operations.
  • This change could make future maintenance and updates more straightforward, reducing the likelihood of bugs.
  • Users can expect the revised code to have similar performance with improved clarity, which might facilitate customization and further development.

@glenn-jocher glenn-jocher self-assigned this Apr 3, 2022
@glenn-jocher glenn-jocher merged commit ad0e4d5 into master Apr 3, 2022
@glenn-jocher glenn-jocher deleted the update/split branch April 3, 2022 18:05
BjarneKuehl pushed a commit to fhkiel-mlaip/yolov5 that referenced this pull request Aug 26, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant