Iterative SE(3)-Transformers

When manipulating three-dimensional data, it is possible to ensure that rotational and translational symmetries are respected by applying so-called SE(3)-equivariant models. Protein structure prediction a prominent example of task which displays these symmetries. Recent work in this area has successfully made use an model, iterative attention mechanism. Motivated application, we implement version the SE(3)-Transformer, attention-based model for graph data. We address additional complications arise when SE(3)-Transformer fashion, compare single-pass versions on toy problem, consider why may be beneficial some problem settings. make code our implementation available community ( https://github.com/FabianFuchsML/se3-transformer-public ).