Hang Su

Can you check if you have the data properly pre-processed as [instructed here](https://github.com/NVlabs/splatnet/blob/master/data/README.md#ruemonge428)? pcl_test.ply should have 11 data columns and 6 should definitely not be out of bounds.

Thanks for the excellent work too! I have another question regarding the speed comparison: you use a [4x4 average pooling](https://github.com/MarvinTeichmann/ConvCRF/blob/master/convcrf/convcrf.py#L350) before doing message passing, effectively reducing the computation 16 folds....

Thanks for the prompt reply! I understand that there is some interpolation happening when mapping the pixels onto a permutohedral lattice. Are you referring to that? There downsampling is done...

Can you check whether the file in question (dependencies/vlfeat/toolbox/vl_setup.m) is indeed pulled properly?

The dependencies folders should be populated with respective libraries after the `git submodule update --init` command. I checked and that should work on both linux and windows.

This is a very common confusion regarding the ModelNet benchmark. There are 12311 shapes in total, but according to the original authors (in their original CVPR paper: '3D ShapeNets: A...

Yes, we fine-tuned the whole network at the end. We actually use three stages of fine-tuning: last layer only, all fc layers, all layers. The number of epochs for each...

The test labels were not provided by the benchmark maintainers: https://shapenet.cs.stanford.edu/shrec16/

I remember seeing about 20% speedup. But that was a while ago and the number might be off...

Sorry for the late reply. Have you figured out a solution? Which file are you referring to?