afe.apis.transform

Source: afe/apis/transform.py

Back to index

Tensor transformations that can be applied to a model's input or output.

Imports

• afe.apis.compilation_job_base.Tensor
• afe.apis.defines.ChromaSampling
• afe.apis.defines.ColorConversion
• afe.apis.defines.ColorSpaceStandard
• afe.apis.defines.ResizeDepositLocation
• afe.apis.defines.ResizeMethod
• afe.backends.Backend
• afe.core.configs.QuantizationConfigs
• afe.ir.defines.NodeName
• afe.ir.tensor_type.ScalarType
• afe.ir.tensor_type.TensorType
• afe.ir.tensor_type.TensorValue
• afe.ir.tensor_type.TupleValue
• afe.ir.tensor_type.data_byte_size
• afe.ir.tensor_type.scalar_type_to_dtype
• dataclasses
• ev_transforms.transforms as ev_transforms
• functools
• numpy as np
• typing.Callable
• typing.Sequence

Classes

• Transform (line 54): A transformation on a tensor.

  • apply(t: Tensor) -> Tensor (line 58): Apply this transform to a tensor. Raise an exception if the tensor's type is not suitable for this transform.

    Parameters:

    • t: type Tensor

    Returns: Tensor

  • get_result_type(parameter_type: TensorType) -> TensorType (line 65): Get the type of this transform's output when it is applied to input of the given type. Raise an exception if the transform cannot be applied to a tensor having the given type.

    Parameters:

    • parameter_type: type TensorType

    Returns: TensorType

  • extract_ir(input_type: TensorType, input_name: NodeName, output_base_name: NodeName) -> tuple[TensorType, list[_AwesomeNode]] (line 73): Extract the internal representation of this transform.

    Parameters:

    • input_type: type TensorType
    • input_name: type NodeName
    • output_base_name: type NodeName

    Returns: tuple[TensorType, list[_AwesomeNode]]

Functions

• argmax(*, axis: int) -> Transform (line 147): The argmax operation applied over a single axis of a tensor.

  Parameters:

  • axis: Axis to reduce

  Returns: The argmax transformation

• reshape_transform(*, newshape: list[int]) -> Transform (line 179): The reshape_transform operation

  Parameters:

  • newshape: New shape after reshape transform

  Returns: The reshape transformation

• layout_transform(*, src_layout: str, dst_layout: str) -> Transform (line 227): The layout_transform operation.

  Parameters:

  • src_layout: Layout of the input tensor
  • dst_layout: Layout of the output tensor

  Returns: The layout_transform transformation

• tessellation_transform(*, slice_shape: Sequence[int], align_c16: bool, cblock: bool) -> Transform (line 314): The tessellation_transform operation.

  Parameters:

  • slice_shape: Shape of slice to tessellate.
  • align_c16: True to align channels to 16 in a tessellated slice
  • cblock: True to interleave the channel blocks in a tessellated slice.

  Returns: The tessellation_transform transformation

• detessellation_transform(*, slice_shape: Sequence[int], align_c16: bool, cblock: bool, frame_type: TensorType) -> Transform (line 395): The detessellation_transform operation.

  Parameters:

  • slice_shape: Shape of slice to detessellate
  • align_c16: True to indicate that slice channels are aligned to 16.
  • cblock: True to indicate that channel blocks are interleaved in a slice.
  • frame_type: Tensor type of de-tessellated frame

  Returns: The detessellation_transform transformation

• pack_transform() -> _PackTransform (line 460): The pack_transform operation.

  Returns: The pack_transform transformation

• unpack_transform(*, tensor_types: list[TensorType]) -> _UnpackTransform (line 516): The unpack_transform operation.

  Parameters:

  • tensor_types: List of target tensor types after unpack

  Returns: The unpack_transform transformation

• normalization_transform(*, channel_params: list[tuple[float, float, float]]) -> Transform (line 561): The normalization_transform operation.

  Parameters:

  • channel_params: The list of tuples for (divisor, mean, standard deviation)

  Returns: The normalization_transform transformation

• quantization_transform(*, channel_params: list[tuple[float, int]], num_bits: int, rounding: _RoundType = _RoundType.TONEAREST) -> Transform (line 622): The quantization_transform operation.

  Parameters:

  • channel_params: The list of tuples for (quant_scale, zero_point)
  • num_bits: The number of bits used for quantization
  • rounding: The rounding type for quantization

  Returns: The quantization_transform transformation

• dequantization_transform(*, channel_params: list[tuple[float, int]]) -> Transform (line 678): The dequantization_transform operation.

  Parameters:

  • channel_params: The list of tuples for (quant_scale, zero_point)

  Returns: The dequantization_transform transformation

• resize_transform(*, target_height: int, target_width: int, keep_aspect: bool, deposit_location: ResizeDepositLocation, method: ResizeMethod) -> Transform (line 730): The resize_transform operation.

  Parameters:

  • target_height: Target height of resized tensor
  • target_width: Target width of resized tensor
  • keep_aspect: Boolean flag to keep aspect ratio
  • deposit_location: Enum to indicate deposit position of resized image
  • method: Enum to indicate supported interpolation methods

  Returns: The resize_transform transformation

• chroma_upsample_transform(*, frame_height: int, frame_width: int, yuv_sampling: ChromaSampling) -> Transform (line 812): The chroma_upsample_transform operation.

  Parameters:

  • frame_height: Height of full sampling frame
  • frame_width: Width of full sampling frame
  • yuv_sampling: Chroma sampling Enum

  Returns: The chroma_upsample_transform transformation

• yuv_rgb_conversion_transform(*, conversion: ColorConversion, std: ColorSpaceStandard) -> Transform (line 869): The yuv_rgb_conversion_transform operation.

  Parameters:

  • conversion: Direction of conversion between YUV and RGB
  • std: Standard for color space conversion

  Returns: The yuv_rgb_conversion_transform transformation

• bgr_rgb_conversion_transform(*, conversion: ColorConversion) -> Transform (line 924): The bgr_rgb_conversion_transform operation.

  Parameters:

  • conversion: Direction of conversion between BGR and RGB

  Returns: The bgr_rgb_conversion_transform transformation

• crop_transform(*, bounding_box: list[tuple[int, int]]) -> Transform (line 1016): The crop_transform operation.

  Parameters:

  • bounding_box: Rectangle area to crop, as a list of 2 items. bounding_box[0] is the (x,y) position of the cropped area's origin within the input area. bounding_box[1] is the (x,y) past-the-end position of the cropped area within the input area.

  Returns: The crop_transform transformation

• slice_transform(*, begin: list[int], end: list[int]) -> Transform (line 1071): The slice_transform operation.

  Parameters:

  • begin: Begin indices for each dimension.
  • end: End indices for each dimension.

  Returns: The slice_transform transformation

• sigmoid_transform() -> Transform (line 1115): The sigmoid_transform operation.

  Returns: The sigmoid_transform transformation

• nms_maxpool_transform(*, kernel: int) -> Transform (line 1157): The nms_maxpool_transform operation.

  Parameters:

  • kernel: Size of pooling kernel

  Returns: The nms_maxpool_transform transformation

• softmax_transform(*, axis: int) -> Transform (line 1192): The softmax transform operation.

  Parameters:

  • axis: Axis to sum over

  Returns: The softmax transform

• compose(transforms: list[Transform]) -> Transform (line 1231): Make a transform that is the composition of the given transforms.

  Parameters:

  • transforms: Transforms to compose

  Returns: Their composition

• identity() -> Transform (line 1272): An identity transformation on a tensor.

  Returns: Transform