stable_learning_control.utils.mpi_utils.mpi_pytorch

Helper methods for managing Pytorch MPI processes.

Note

This module is not yet used in any of the current algorithms, but is kept here for future reference.

Module Contents

Functions

setup_pytorch_for_mpi()	Avoid slowdowns caused by each separate process's PyTorch using more than its
mpi_avg_grads(module)	Average contents of gradient buffers across MPI processes.
sync_params(module)	Sync all parameters of module across all MPI processes.

stable_learning_control.utils.mpi_utils.mpi_pytorch.setup_pytorch_for_mpi()

Avoid slowdowns caused by each separate process’s PyTorch using more than its fair share of CPU resources.

stable_learning_control.utils.mpi_utils.mpi_pytorch.mpi_avg_grads(module)

Average contents of gradient buffers across MPI processes.

Parameters:

module (object) – Python object for which you want to average the gradients.

stable_learning_control.utils.mpi_utils.mpi_pytorch.sync_params(module)

Sync all parameters of module across all MPI processes.

Parameters:

module (object) – Python object for which you want to average the gradients.