d3rlpy.preprocessing.PixelScaler¶

class d3rlpy.preprocessing.PixelScaler[source]¶

Pixel normalization preprocessing.

\[x' = x / 255\]

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

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

# initialize algorithm with PixelScaler
cql = CQL(scaler='pixel')

cql.fit(dataset.episodes)

Methods

fit(episodes)[source]¶

get_params()[source]¶

Returns scaling parameters.

PixelScaler returns empty dictiornary.

Returns:	empty dictionary.
Return type:	dict

get_type()[source]¶

Returns scaler type.

Returns:	pixel.
Return type:	str

reverse_transform(x)[source]¶

Returns reversely transformed observations.

Parameters:	x (torch.Tensor) – normalized observation tensor.
Returns:	unnormalized pixel observation tensor.
Return type:	torch.Tensor

transform(x)[source]¶

Returns normalized pixel observations.

Parameters:	x (torch.Tensor) – pixel observation tensor.
Returns:	normalized pixel observation tensor.
Return type:	torch.Tensor