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 was wondering if someone could kindly explain to me the reasoning behind the specific weight values assigned to each loss in the code. I notice that the loss of weight is :
" normal_loss_weight: 0.1
edge_loss_weight: 20
joint_loss_weight: 0.001
"
I want to know how to determine each weight of the loss.
Why is the normal_loss_weight set to 0.1, the edge_loss_weight set to 20, and the joint_loss_weight set to 0.001?
Thank you for your time!
The text was updated successfully, but these errors were encountered:
The joint_loss_weight was fixed to 0.001, since the joint coordinates are in mm scale. The edge loss was important to get smooth mesh surfaces. However, initially giving the edge loss entailed trivial solutions, so we gave the loss after some training.
Hello, I was wondering if someone could kindly explain to me the reasoning behind the specific weight values assigned to each loss in the code. I notice that the loss of weight is :
" normal_loss_weight: 0.1
edge_loss_weight: 20
joint_loss_weight: 0.001
"
I want to know how to determine each weight of the loss.
Why is the normal_loss_weight set to 0.1, the edge_loss_weight set to 20, and the joint_loss_weight set to 0.001?
Thank you for your time!
The text was updated successfully, but these errors were encountered: