ERBS is a library that provides enhanced sampling methods for generating Machine Learning Potential Training data.
It is build on top of Jax and can be used with any atomistic descriptor that matches our function signature. The library itself only ships with a simple example descriptor, mostly for testing reasons. We recommend using our other package, APAX, to access the more powerful gaussian moment descriptor.