20 lines (16 loc) · 433 Bytes

.. role:: hidden
    :class: hidden-section

Distributed Optimizers

:::{warning} Distributed optimizer is not currently supported when using CUDA tensors :::

.. automodule:: torch.distributed.optim
    :members: DistributedOptimizer, PostLocalSGDOptimizer, ZeroRedundancyOptimizer

.. currentmodule:: torch.distributed.optim.utils
.. autofunction:: register_functional_optim