Super parameter to ShapeNet car dataset

[Gardlin]

hello, when I reproduce the pre-training experiment on your shapenet car data set, I found that the training does not converge very well. I think it is related to the parameter of piece-wise_constant, wish you can provide the hyperparameter setting?

[hansen7]

Sorry which part you are referring to? I can't remember which is the "shapenet car data set".

[Gardlin]

https://drive.google.com/drive/folders/13v9To7ynkhWhC0CN1X7gJehRPfNZREBU
I mean how to set alpha for the above dataset , and do you provide the trained model for this dataset

[hansen7]

Thanks, I haven't worked on this dataset, you might contact the author team of Point Completion Network :)

[Gardlin]

By the way , do you provide the well-trained models on ModelNet, I followed your hyper parameter and train 50 epochs ,however it seems that i can't get a good visualization result.

[hansen7]

would you mind elaborating a bit more? the visualisation can be reproduced on the provided pre-trained models, it is in the "Pre-Trained Weights" section in the readme file :)

[Gardlin]

Thanks, I mean when I run python train_completion.py on the modelnet dataset(ShapeNet car Dataset) ，the completion didnot seem to perform well. I checked the provided pre-trained models,but there are not the pointnet_occo pretrained models for point cloud completion.