Clears the state of the datum.
Clears the state of the datum.
Clears the state of the model.
Clears the state of the model.
Creates a new datum for processing with the given model.
Creates a new datum for processing with the given model.
Creates a new model from the given model parameters.
Creates a new model from the given model parameters.
Collects ModelStates into a new summary model state.
Collects ModelStates into a new summary model state.
Does one iteration of learning on the given data, updating ModelState and each Datum's DatumState as appropriate.
Does one iteration of learning on the given data, updating ModelState and each Datum's DatumState as appropriate.
Gets the state of the datum.
Gets the state of the datum.
Get model parameters used to construct a model like the given.
Get model parameters used to construct a model like the given.
Gets the state of the model.
Gets the state of the model.
Uses the given datum state.
Uses the given datum state.
Uses the given model state.
Uses the given model state.
DataParallel models can do assignments in parallel, updating model parameters locally, aggregating model parameters upwards, e.g. in a tree sum.