Posts by Tres Popp
DGL in Depth: SE(3)-Transformer on ROCm 7
- 05 December 2025
In this post, we demonstrate how to run the SE(3)-Transformer efficiently with Deep Graph Library (DGL) on AMD ROCm, enabling high-performance 3D graph learning for complex geometric models. This builds on our previous blog, which highlighted DGL’s versatility across diverse graph neural network (GNN) workloads, validating functionality, compatibility, and usability.