simrun ❭ modular_reduced_model_inference ❭ solver ❭ _Solver ❭ optimize_one_split

_Solver.optimize_one_split

simrun.modular_reduced_model_inference.solver._Solver.optimize_one_split(client=None, workers=None, index=0)

Optimize the cost function for a single split.

Parameters:

  • client (dask.distributed.Client) – A dask client object.

  • workers (list) – List of worker names. Passed to dask.distributed.Client.submit()

  • index (int) – Index of the split to optimize. Splits are saved as a dictionary in the splits attribute. The index then refers to the index of the names (i.e. keys) of the splits. Note that dictionaries are in general unordered, so this index is only useful to differentiate between splits, not to refer to a specific split or order of splits.

Returns:

A dictionary with the optimization results.

Return type:

dict