hippynn
hippynn copied to clipboard
lanl
python library for atomistic machine learning
PyTorch is depreciating the argument 'verbose' in the ReduceLROnPlateau function. The HIPPYNN function RaiseBatchSizeOnPlateau will need to be updated to avoid using this argument when calling ReduceLROnPlateau.
For reference: ```python # Assuming fij and atom_energy are tensors and data is an object with an eatoms attribute if return_device == "cpu": # Convert tensors to numpy arrays if...
Using hippynn via mliap causes script to fail: ``` Setting up Verlet run ... Unit style : metal Current step : 0 Time step : 0.001 Traceback (most recent call...
@yingwaili requests that we do the training in such away that restarting the model is basically automated. Maybe we can have some function to write a restart script in the...
Also includes minor changes to examples
ASE interface supports different units, we need to add the capability to the LAMMPS interface as well.
@bnebgen-LANL I will merge this when you test and let me know that it doesn't disrupt your workflows either on kokkos or non-kokkos.
Right now there are functions for finding/making different node types in different places, such as acquire_padding_encoding, batch geometry optimizers, ase setup, lammps setup. Some parts of these functionalities are really...
I am interested in using hippynn to do charge modeling, is there any working examples of training hip-nn to charges or dipole moments as in the paper "Discovering a Transferable...
