Accuracy is low for Validation dataset

drprojects / DeepViewAgg

[CVPR'22 Best Paper Finalist] Official PyTorch implementation of the method presented in "Learning Multi-View Aggregation In the Wild for Large-Scale 3D Semantic Segmentation"

Other

222 stars 24 forks source link

Accuracy is low for Validation dataset #37

Closed adikrishnanucd closed 1 year ago

adikrishnanucd commented 1 year ago

Hello @drprojects,

I was wondering if there are any changes I need to make to run the kitti360_inference code with the CPU instead of the GPU.

Thank you, Aditya

drprojects commented 1 year ago

Hi, we do not support CPU inference, because torchsprase does not. If you wanted to run on CPU, you would need to switch to MinkowskiEngine sparse convolution library, but I do not have time to provide support for this. Best, Damien