ma-xu/pointMLP-pytorch

Why is there only 9480 ply_data_train in modelnet40 dataset instead of 9483

Closed this issue · 5 comments

Why is there only 9480 ply_data_train in modelnet40 dataset instead of 9483
ma-xu commented

It may caused by different released versions of the dataset (not sure).

Please refer the official dataset link:

www = 'https://shapenet.cs.stanford.edu/media/modelnet40_ply_hdf5_2048.zip'

Training set really only has 9480 data in ’modelnet40_ply_hdf5_2048’ .

Will these three data have an effect on the results of the experiment?

ma-xu commented

Will these three data have an effect on the results of the experiment?

Not sure if different verions will have some performance fluctuations. All our experiments are conducted on this version (as well as CurveNet, PAConv, GDANet, etc.)

Will these three data have an effect on the results of the experiment?

Not sure if different verions will have some performance fluctuations. All our experiments are conducted on this version (as well as CurveNet, PAConv, GDANet, etc.)

Thank you for your reply, thanks!