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
Hi!
I am using solov2 to train my dataset, but the size of the images in my dataset is different from coco dataset. The aspect ratio of the images is 3:4 or 1:1 (I resize the original dimensions and scale to the image scale(height = 1080)). So how can I adjust num_grids and scale_ranges to get better AP. Can you give some suggestions? For example how can I calculate to get num_grids which can make model perform better, whether can I reset scale_ranges value to get better result or not?
Thank you! @WXinlong
The text was updated successfully, but these errors were encountered:
Hi!
I am using solov2 to train my dataset, but the size of the images in my dataset is different from coco dataset. The aspect ratio of the images is 3:4 or 1:1 (I resize the original dimensions and scale to the image scale(height = 1080)). So how can I adjust num_grids and scale_ranges to get better AP. Can you give some suggestions? For example how can I calculate to get num_grids which can make model perform better, whether can I reset scale_ranges value to get better result or not?
Thank you!
@WXinlong
The text was updated successfully, but these errors were encountered: