ray.rllib.algorithms.algorithm.Algorithm.restore

Algorithm.restore(checkpoint_path: str | Checkpoint | _TrainingResult)

Restores training state from a given model checkpoint.

These checkpoints are returned from calls to save().

Subclasses should override load_checkpoint() instead to restore state. This method restores additional metadata saved with the checkpoint.

checkpoint_path should match with the return from save().

Parameters:

checkpoint_path – training result that was returned by a previous call to save().

DeveloperAPI: This API may change across minor Ray releases.