Use the following instructions to launch a short training job on QM9/H. See default.py for the longer configurations that reproduce the results in the paper. It should train at around 21.9 steps/s ...
This work is a significant extension of our original work presented in IEEE CVPR2019, and is accepted to TPAMI in March 2020. We propose a spherical kernel for efficient graph convolution of 3D point ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback