d3rlpy.augmentation.image.ColorJitter¶

class d3rlpy.augmentation.image.ColorJitter(brightness=(0.6, 1.4), contrast=(0.6, 1.4), saturation=(0.6, 1.4), hue=(- 0.5, 0.5))[source]¶

Color Jitter augmentation.

This augmentation modifies the given images in the HSV channel spaces as well as a contrast change. This augmentation will be useful with the real world images.

References

Laskin et al., Reinforcement Learning with Augmented Data.

Parameters

brightness (tuple) – brightness scale range.
contrast (tuple) – contrast scale range.
saturation (tuple) – saturation scale range.
hue (tuple) – hue scale range.

Methods

get_params(deep=False)[source]¶

Returns augmentation parameters.

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

get_type()¶

Returns augmentation type.

Returns: augmentation type.
Return type: str

transform(x)[source]¶

Returns augmented observation.

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

Attributes

TYPE: ClassVar[str] = 'color_jitter'¶