transforms.Resize¶

class lucid.transforms.Resize(size: tuple[int, int])¶

The Resize class is a transformation that resizes an input tensor to the specified dimensions. This transformation is useful for ensuring that input tensors have a consistent size, which is often required in image processing and neural network applications.

Class Signature¶

class lucid.transforms.Resize(size: tuple[int, int])

Parameters¶

size (tuple[int, int]): The target size to which the input tensor will be resized. The tuple represents the height and width of the output tensor.

Attributes¶

size (tuple[int, int]): The size to which the input tensor will be resized, stored as an internal attribute.

Forward Calculation¶

The resizing of an input tensor \(\mathbf{x}\) is performed using interpolation or down-sampling techniques, depending on the relationship between the original size and the target size.

If the target size is larger than the original size, up-sampling is performed.

If the target size is smaller than the original size, down-sampling is performed.

The transformation is applied independently for each channel.

Examples¶

Example 1: Basic Usage

>>> import lucid.transforms as T
>>> input_tensor = lucid.Tensor([[[1.0, 2.0], [3.0, 4.0]]])  # Shape: (1, 2, 2)
>>> resize = T.Resize(size=(4, 4))  # Resize to 4x4
>>> output = resize(input_tensor)
>>> print(output)
Tensor([[[1.0, 1.5, 2.0, 2.0],
         [2.0, 2.5, 3.0, 3.0],
         [3.0, 3.5, 4.0, 4.0],
         [3.0, 3.5, 4.0, 4.0]]], grad=None)

Example 2: Resizing an Image Batch

>>> import lucid.transforms as T
>>> input_tensor = lucid.Tensor([[[[1.0, 2.0], [3.0, 4.0]], [[5.0, 6.0], [7.0, 8.0]]]])  # Shape: (1, 2, 2, 2)
>>> resize = T.Resize(size=(4, 4))  # Resize each image in the batch to 4x4
>>> output = resize(input_tensor)
>>> print(output)
Tensor([[[[1.0, 1.5, 2.0, 2.0],
          [2.0, 2.5, 3.0, 3.0],
          [3.0, 3.5, 4.0, 4.0],
          [3.0, 3.5, 4.0, 4.0]],

         [[5.0, 5.5, 6.0, 6.0],
          [6.0, 6.5, 7.0, 7.0],
          [7.0, 7.5, 8.0, 8.0],
          [7.0, 7.5, 8.0, 8.0]]]], grad=None)

Note

The resize operation maintains the aspect ratio if both dimensions are scaled proportionally.
This transformation is typically applied to images to ensure a consistent input size for neural networks.