DefRec_and_PCM icon indicating copy to clipboard operation
DefRec_and_PCM copied to clipboard

Published 20 hours ago • verified publisher icon IdanAchituve

About

Open lsssssssssss opened this issue 3 years ago • 4 comments

I couldn't get the visualizations I wanted after the reconstruction.The network doesn't seem to be converging.Could you please send me the weight parameters after pre-training

lsssssssssss avatar Oct 11 '21 09:10 lsssssssssss

Hi, I didn't experience any convergence issues. Which variant of the model are you running? What diverges exactly? I don't have the parameters of the learned model.

IdanAchituve avatar Oct 11 '21 10:10 IdanAchituve

When I visualized the reconstructed point cloud, I got a jumble of points

lsssssssssss avatar Oct 11 '21 10:10 lsssssssssss

How can I get an effect like figure 9-11

lsssssssssss avatar Oct 11 '21 10:10 lsssssssssss

As seen in the paper, the reconstruction is not very good as this isn't the main point of the method. However, if the training goes well and the reconstruction loss decreases during training the plots should come out reasonable. I am not sure why you got the results that you did.

IdanAchituve avatar Oct 11 '21 12:10 IdanAchituve