GradDFT is a JAX-based library enabling the differentiable design and experimentation of exchange-correlation functionals using machine learning techniques.
Apache License 2.0
79
stars
7
forks
source link
Unifying the notation between the jitted and non-jitted orbital optimizers. #74