helios.optim.utils¶

Attributes¶

OPTIMIZER_REGISTRY

Global instance of the registry for optimizers.

Functions¶

create_optimizer(→ torch.nn.Module)

Create the optimizer for the given type.

Module Contents¶

helios.optim.utils.OPTIMIZER_REGISTRY¶

Global instance of the registry for optimizers.

By default, the registry contains the following optimizers:

Optimizers¶
Optimizer	Name
`torch.optim.Adadelta`	Adadelta
`torch.optim.Adagrad`	Adagrad
`torch.optim.Adam`	Adam
`torch.optim.AdamW`	AdamW
`torch.optim.SparseAdam`	SparseAdam
`torch.optim.Adamax`	Adamax
`torch.optim.ASGD`	ASGD
`torch.optim.LBFGS`	LBFGS
`torch.optim.NAdam`	NAdam
`torch.optim.RAdam`	RAdam
`torch.optim.RMSprop`	RMSprop
`torch.optim.Rprop`	Rprop
`torch.optim.SGD`	SGD

Example

import helios.optim as hlo

# This automatically registers your optimizer.
@hlo.OPTIMIZER_REGISTRY.register
class MyOptimizer:
    ...

# Alternatively you can manually register a optimizer like this:
hlo.OPTIMIZER_REGISTRY.register(MyOptimizer)

helios.optim.utils.create_optimizer(type_name: str, *args: Any, **kwargs: Any) → torch.nn.Module¶

Create the optimizer for the given type.

Parameters:

type_name – the type of the optimizer to create.
args – positional arguments to pass into the optimizer.
kwargs – keyword arguments to pass into the optimizer.

Returns:

The optimizer.