MultiMadnisTraining

class madspace.MultiMadnisTraining(self: madspace._madspace_py.MultiMadnisTraining, generator_context: madspace._madspace_py.Context, optimizer_context: madspace._madspace_py.Context, training_args: collections.abc.Sequence[madspace._madspace_py.TrainingArgs], verbosity: madspace._madspace_py.Verbosity, status_file: madspace._madspace_py.StatusFile = None, seed: SupportsInt | None = None)

Bases: pybind11_object

Runs MadnisTraining for every subprocess of a process, with combined progress reporting.

Parameters:

  • generator_context – Context the sampling runtimes run on.

  • optimizer_context – Context the trained globals and optimizers live on.

  • training_args – One entry per subprocess.

  • verbosity – Progress-reporting verbosity.

  • status_file – Optional status file progress is periodically written to.

  • seed – Top-level run seed (reusing the run card’s own seed, also used by build_event_generator()); each subprocess’s MadnisTraining gets a channel_index_offset slice of this seed’s stream so their derived seeds don’t collide.

active_channels(self: madspace._madspace_py.MultiMadnisTraining) → list[list[int]]

Active channel indices of every subprocess; see MadnisTraining::active_channels.

train(self: madspace._madspace_py.MultiMadnisTraining) → None

Train every subprocess to completion.