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
Thanks for the good research. I'm trying to apply FeatUp to an object detection task. Will there be any performance improvement? Specifically I'm using the yolo series.
When upscaling the feature map in the pyramid space, applying FeatUP to create a more refined feature map seems to help improve performance. What do you think?
Is custom training needed for this? Or can I use pre-trained weights?
The text was updated successfully, but these errors were encountered:
Thanks for the good research. I'm trying to apply FeatUp to an object detection task. Will there be any performance improvement? Specifically I'm using the yolo series.
When upscaling the feature map in the pyramid space, applying FeatUP to create a more refined feature map seems to help improve performance. What do you think?
Is custom training needed for this? Or can I use pre-trained weights?
The text was updated successfully, but these errors were encountered: