search

atomistic-machine-learning / schnetpack

SchNetPack - Deep Neural Networks for Atomistic Systems
Other
791 stars 215 forks source link

Emphasize where strain needs to be in inputs #673

Closed WardLT closed 1 week ago

WardLT commented 1 week ago

Took me a long time to figure out why the strain was not being used when computing forces. Felt stupid afterwards, but also that it could be an easy mistake others would make.

WardLT commented 1 week ago

For clarity, I was getting a RuntimeError: One of the differentiated Tensors appears to not have been used in the graph. Set allow_unused=True if this is the desired behavior. when using calc_stress=True for the Forces output module.

jnsLs commented 1 week ago

Thank you for your contribution!