def run

stelliferous@0.0.2.0 · kernel_weights.bend:70

import Base
import stelliferous@0.0.2.0/storage_allocation.bend as Allocation
import stelliferous@0.0.2.0/traversal_array.bend as Traversal
import stelliferous@0.0.2.0/kernel_tile.bend as Tile
import stelliferous@0.0.2.0/tensor.bend as Tensors

70def run(+rows: U32,+inner: U32,input: Array<F32>) -> Array<Tile.Tile8>:
71  +blocks = ((rows + 7) / 8 : U32)
72  reordered(Traversal.indexed_fold(~(Array<F32> & Array<Tile.Tile8>),~Layout,~copy_block,U32.to_nat(blocks),0,
73    Layout{rows,inner},(input,Allocation.array(~Tile.Tile8,Tensors.depth_for((blocks * inner : U32)),Tile.filled(0.0)))))