kimren227 / extrudenet Goto Github PK

I'm traing your code on category plane,it shows that training needs 10000 epoches,and the model will save if current test loss is lower than before,i want wo know how many epoch is enough for single category in your ex.Thank you very much!!!

Hi,

Thanks for sharing the code. I am wondering if it is possible for you to also release the pretrained checkpoint for ShapeNet planes and chairs?

Thanks

Hi,

Thank you for your great work on shape parsing and CAD reconstruction of 3D point cloud.
I and my colleagues recently came across your paper and got really inspired by it, and trying to build upon your brilliant work.

I'm trying to train a model with preprocessed chair dataset with the given chair config file, however the key error of loss dictionary happens during the training.
In the training code, 'Loss_Drift', 'Loss_Support', 'Loss_Control_Polygon' are supposed to be included in the loss dictionary. But the Loss class in loss.py doesn't have these losses.
So I'm running the code without these losses by commenting them out, but I'm wondering that it's okay.

If I'm missing something here, please let me know.

Thank you in advance.
Eunji.