You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Recently, I conducted experiments comparing different models and found that using transformer-based models for rotation object detection did not perform well. I recall reading that the QKV mechanism might not be well-suited for objects with large scale variations. Could anyone share more detailed papers or explanations on the disadvantages of QKV in the application of rotation object detection?
Any other context?
No response
The text was updated successfully, but these errors were encountered:
What's the feature?
Recently, I conducted experiments comparing different models and found that using transformer-based models for rotation object detection did not perform well. I recall reading that the QKV mechanism might not be well-suited for objects with large scale variations. Could anyone share more detailed papers or explanations on the disadvantages of QKV in the application of rotation object detection?
Any other context?
No response
The text was updated successfully, but these errors were encountered: