Mojo function

roi_align_nhwc

roi_align_nhwc[type: DType, output_layout: Layout, input_layout: Layout, roi_layout: Layout, //, aligned: Bool, mode: StringSlice[StaticConstantOrigin] = __init__[__mlir_type.!kgen.string]("AVG")](output: LayoutTensor[type, output_layout, origin, address_space=address_space, element_layout=element_layout, layout_int_type=layout_int_type, linear_idx_type=linear_idx_type, masked=masked, alignment=alignment], input: LayoutTensor[type, input_layout, origin, address_space=address_space, element_layout=element_layout, layout_int_type=layout_int_type, linear_idx_type=linear_idx_type, masked=masked, alignment=alignment], rois: LayoutTensor[type, roi_layout, origin, address_space=address_space, element_layout=element_layout, layout_int_type=layout_int_type, linear_idx_type=linear_idx_type, masked=masked, alignment=alignment], output_height: Int, output_width: Int, in_spatial_scale: SIMD[dtype, 1], in_sampling_ratio: SIMD[dtype, 1])

Compute ROIAlign a batch of rois of shape [M, 5] where the first dim is the batch index, followed by region box coordinates (y0, x0) (y1, x1). For inputs of NHWC format. The output shape is [M, output_height, output_width, C].

Parameters:

type (DType): Type of the input tensor.
output_layout (Layout): The output layout.
input_layout (Layout): The input layout.
roi_layout (Layout): The layout of the regions of interests (ROI).
aligned (Bool): If not true offset the ROIs by 0.5.
mode (StringSlice[StaticConstantOrigin]): The pooling mode "AVG" for average and "MAX" for max pooling.