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
Hello, I would like to fine-tune the metric depth network on my own outdoor dataset. Can I use depth_anything_v2_metric_vkitti_vitl.pth as the pre-trained model to start training instead of depth_anything_v2_vitl.pth? Would this approach yield better results?
The text was updated successfully, but these errors were encountered:
I think you can try both strategies. But personally, I would recommend using depth_anything_v2_vitl.pth as a pre-training source, since it is trained on much larger-scale data and is more general.
Hello, I would like to fine-tune the metric depth network on my own outdoor dataset. Can I use depth_anything_v2_metric_vkitti_vitl.pth as the pre-trained model to start training instead of depth_anything_v2_vitl.pth? Would this approach yield better results?
The text was updated successfully, but these errors were encountered: