d3rlpy.preprocessing.StandardScaler¶

class d3rlpy.preprocessing.StandardScaler(dataset=None, mean=None, std=None, eps=0.001)[source]¶

Standardization preprocessing.

\[x' = (x - \mu) / \sigma\]

from d3rlpy.dataset import MDPDataset
from d3rlpy.algos import CQL

dataset = MDPDataset(observations, actions, rewards, terminals)

# initialize algorithm with StandardScaler
cql = CQL(scaler='standard')

# scaler is initialized from the given episodes
cql.fit(dataset.episodes)

You can initialize with d3rlpy.dataset.MDPDataset object or manually.

from d3rlpy.preprocessing import StandardScaler

# initialize with dataset
scaler = StandardScaler(dataset)

# initialize manually
mean = observations.mean(axis=0)
std = observations.std(axis=0)
scaler = StandardScaler(mean=mean, std=std)

cql = CQL(scaler=scaler)

Parameters

dataset (d3rlpy.dataset.MDPDataset) – dataset object.
mean (numpy.ndarray) – mean values at each entry.
std (numpy.ndarray) – standard deviation at each entry.
eps (float) – small constant value to avoid zero-division.

Methods

fit(episodes)[source]¶

Estimates scaling parameters from dataset.

Parameters: episodes (List[d3rlpy.dataset.Episode]) – list of episodes.
Return type: None

fit_with_env(env)[source]¶

Gets scaling parameters from environment.

Parameters: env (gym.core.Env) – gym environment.
Return type: None

get_params(deep=False)[source]¶

Returns scaling parameters.

Parameters: deep (bool) – flag to deeply copy objects.
Returns: scaler parameters.
Return type: Dict[str, Any]

get_type()¶

Returns a scaler type.

Returns: scaler type.
Return type: str

reverse_transform(x)[source]¶

Returns reversely transformed observations.

Parameters: x (torch.Tensor) – observation.
Returns: reversely transformed observation.
Return type: torch.Tensor

transform(x)[source]¶

Returns processed observations.

Parameters: x (torch.Tensor) – observation.
Returns: processed observation.
Return type: torch.Tensor

Attributes

TYPE: ClassVar[str] = 'standard'¶