For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /max/get-started.md).

Mojo function

pad_reflect

def pad_reflect[dtype: DType, paddings_type: DType](output: TileTensor[dtype, Storage=output.Storage, linear_idx_type=output.linear_idx_type], input: TileTensor[dtype, Storage=input.Storage, linear_idx_type=input.linear_idx_type], paddings: Pointer[Scalar[paddings_type], _safe=False])

Fill output with values from input, and edges padded with reflected values from the unpadded region.

Example: var input = [[1, 2], [3, 4]] var paddings = [2, 2, 1, 0]

Yields: output = [[2, 1, 2], [4, 3, 4], [2, 1, 2], [4, 3, 4], [2, 1, 2], [4, 3, 4]]

Parameters:

dtype (DType): DType of the input and output buffers.
paddings_type (DType): DType of the paddings buffer.

Args:

output (TileTensor[dtype, Storage=output.Storage, linear_idx_type=output.linear_idx_type]): The output buffer.
input (TileTensor[dtype, Storage=input.Storage, linear_idx_type=input.linear_idx_type]): The input buffer.
paddings (Pointer[Scalar[paddings_type], _safe=False]): Ordered (before, after) padding sizes for each axis.