torchcvnn.transforms.HWC2CHW

class torchcvnn.transforms.HWC2CHW(dtype: str | None = None)

Converts an input array/tensor from HWC to CHW format.

This transform reorders the dimensions of the input array/tensor from Height x Width x Channel to Channel x Height x Width format.

Returns:

Input with dimensions reordered to Channel x Height x Width.

Return type:

np.ndarray | torch.Tensor

Example

>>> transform = HWC2CHW()
>>> x = torch.randn(3, 4, 5)  # Shape (3, 4, 5)
>>> y = transform(x)          # Shape (5, 3, 4)

__init__(dtype: str | None = None) → None

Methods

__init__([dtype])