xpag.agents.rljax_agents.util.optim#

Functions

clip_gradient

Clip gradients.

clip_gradient_norm

Clip norms of gradients.

optimize

rtype:

Tuple[Any, Mapping[str, Mapping[str, Array]], Array, Any]

soft_update

Update target network using Polyak-Ruppert Averaging.

weight_decay

Calculate the sum of L2 norms of parameters.