d3rlpy.preprocessing.MinMaxScaler¶
-
class
d3rlpy.preprocessing.
MinMaxScaler
(dataset=None, maximum=None, minimum=None)[source]¶ Min-Max normalization preprocessing.
\[x' = (x - \min{x}) / (\max{x} - \min{x})\]from d3rlpy.dataset import MDPDataset from d3rlpy.algos import CQL dataset = MDPDataset(observations, actions, rewards, terminals) # initialize algorithm with MinMaxScaler cql = CQL(scaler='min_max') # scaler is initialized from the given episodes cql.fit(dataset.episodes)
You can also initialize with
d3rlpy.dataset.MDPDataset
object or manually.from d3rlpy.preprocessing import MinMaxScaler # initialize with dataset scaler = MinMaxScaler(dataset) # initialize manually minimum = observations.min(axis=0) maximum = observations.max(axis=0) scaler = MinMaxScaler(minimum=minimum, maximum=maximum) cql = CQL(scaler=scaler)
- Parameters
dataset (d3rlpy.dataset.MDPDataset) – dataset object.
min (numpy.ndarray) – minimum values at each entry.
max (numpy.ndarray) – maximum values at each entry.
Methods
-
fit
(episodes)[source]¶ Fits minimum and maximum from list of episodes.
- Parameters
episodes (List[d3rlpy.dataset.Episode]) – list of episodes.
- Return type
-
reverse_transform
(x)[source]¶ Returns reversely transformed observations.
- Parameters
x (torch.Tensor) – normalized observation tensor.
- Returns
unnormalized observation tensor.
- Return type
torch.Tensor
-
transform
(x)[source]¶ Returns normalized observation tensor.
- Parameters
x (torch.Tensor) – observation tensor.
- Returns
normalized observation tensor.
- Return type
torch.Tensor
Attributes