RewardNormalizer

class torchrl.trainers.RewardNormalizer(decay: float = 0.999, scale: float = 1.0, eps: float | None = None, log_pbar: bool = False, reward_key=None)

獎勵歸一化器鉤子。

引數:

• decay (float, 可選) – 指數移動平均衰減引數。預設為 0.999

• scale (float, 可選) – 歸一化後的獎勵用於相乘的縮放因子。預設為 1.0。

• eps (float, 可選) – 用於防止數值下溢的 epsilon 抖動。預設為 torch.finfo(DEFAULT_DTYPE).eps，其中 DEFAULT_DTYPE=torch.get_default_dtype()。

• reward_key (str 或 tuple, 可選) – 在輸入批次中查詢獎勵的鍵。預設為 ("next", "reward")

示例

>>> reward_normalizer = RewardNormalizer()
>>> trainer.register_op("batch_process", reward_normalizer.update_reward_stats)
>>> trainer.register_op("process_optim_batch", reward_normalizer.normalize_reward)

register(trainer: Trainer, name: str = 'reward_normalizer')

Registers the hook in the trainer at a default location.

引數:

• trainer (Trainer) – the trainer where the hook must be registered.

• name (str) – the name of the hook.

注意

To register the hook at another location than the default, use register_op().