Why use the anchors clustered on Coco dataset to train VOC?
apxlwl opened this issue · 1 comments
apxlwl commented
Thanks for your excellent work. But a little question here. The default anchor sizes are clustered on Coco dataset according to the paper. Do they also work well for the pascal dataset? Why don't you cluster on the pascal dataset to get a group of specific anchors?
mileistone commented
it doesn't matter