Add "Export to executorch" to Ultralytics HUB

[agoransson]

Hello, it would be very good to be able to easily export to executorch from the Ultralytics HUB.

https://pytorch.org/executorch/stable/tutorials/export-to-executorch-tutorial.html

Thanks.

[UltralyticsAssistant]

👋 Hello @agoransson, thank you for your suggestion and for being part of the Ultralytics community 🚀! We love hearing feedback and ideas to improve the experience for our users.

If this feature request relates to any current limitations or specific use cases, feel free to provide additional details to help guide the team. The PyTorch Export-to-Executorch Tutorial you referenced is a great start, and we appreciate you pointing this out!

For now, we recommend ensuring you're running the latest ultralytics package in case any updates or export-related functionalities are already available. You can upgrade with:

pip install -U ultralytics

Additionally, if you're already experimenting with custom export options, we'd love to hear about your findings or challenges. The Ultralytics HUB is still evolving, and any insights you can share will be invaluable 😊.

If you’re looking for more resources:

• Check our Docs for additional guides related to model exporting.
• Join the Ultralytics community for real-time discussions on Discord 🎧, topic threads on Discourse, and collaboration on Subreddit.

Rest assured, this is an automated response 🤖, but our engineering team will review your request and provide further input soon. Thank you for contributing to making Ultralytics even better! 🚀

[glenn-jocher]

You can export YOLO models to various formats, including TorchScript, which might serve as a starting point for your Executorch needs, and you can explore other export options like ONNX or TensorFlow Lite as described in the Ultralytics documentation on Export mode.