You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
For some QCFractal OptimizationDataset workflows, it would be very useful to take at most K optimization steps and still return successful completion. For example, when starting from snapshots sampled from an MM surrogate potential at 300K, the initial QM energies may be too high, but a few steps of optimization is sufficient to explore the thermodynamically relevant energy surface in the vicinity of the starting geometry.
QCFractal / QCEngine appears to call geomeTRIC in a manner that passes in a JSON document via run_json. Options such as whether to use tric or cart appear to be passed this way. If an option such as success_on_maxiter could be passed to signal that reaching maxiter should still successfully return completion of the optimization trajectory, this use case could easily be supported.
The text was updated successfully, but these errors were encountered:
For some QCFractal OptimizationDataset workflows, it would be very useful to take at most K optimization steps and still return successful completion. For example, when starting from snapshots sampled from an MM surrogate potential at 300K, the initial QM energies may be too high, but a few steps of optimization is sufficient to explore the thermodynamically relevant energy surface in the vicinity of the starting geometry.
QCFractal / QCEngine appears to call geomeTRIC in a manner that passes in a JSON document via run_json. Options such as whether to use
tric
orcart
appear to be passed this way. If an option such assuccess_on_maxiter
could be passed to signal that reachingmaxiter
should still successfully return completion of the optimization trajectory, this use case could easily be supported.The text was updated successfully, but these errors were encountered: