def new_spec

bendtorch-core@0.1.0.0 · src/autograd/reduce_replayed_proof.bend:277

import Base
import bendtorch-core@0.1.0.0/src/shape/core.bend as Shape
import bendtorch-core@0.1.0.0/src/storage/model.bend as Model
import bendtorch-core@0.1.0.0/src/storage/LAWS.bend as StorageLaws

277def new_spec(+plan: Shape.BroadcastPlan, +map: PlanMap(plan), +shape: Shape.Shape, +fill: F32, +ops: List<&2, StorageLaws.Op>, +depth: Nat, +base: Nat) -> {Array.new(F32, depth, 0.0) == out_spec(plan, map, shape, fill, ops, 0n, depth, base) : Array<F32>}:
278  match depth:
279    case 0n:
280      %Equal.sym(Bool, Nat.is_lt(base, 0n), False{}, lt_zero_false(base)) :
281        {ALeaf{0.0} == ALeaf{out_leaf(plan, map, shape, fill, ops, base, _)} : Array<F32>}
282      {==}
283    case 1n+inner:
284      %Equal.sym(Array<F32>, Array.new(F32, inner, 0.0), out_spec(plan, map, shape, fill, ops, 0n, inner, base), new_spec(plan, map, shape, fill, ops, inner, base)) :
285        {ANode{_, Array.new(F32, inner, 0.0)} == out_spec(plan, map, shape, fill, ops, 0n, 1n+inner, base) : Array<F32>}
286      %Equal.sym(Array<F32>, Array.new(F32, inner, 0.0), out_spec(plan, map, shape, fill, ops, 0n, inner, Nat.add(Model.capacity(inner), base)),
287        new_spec(plan, map, shape, fill, ops, inner, Nat.add(Model.capacity(inner), base))) :
288        {ANode{out_spec(plan, map, shape, fill, ops, 0n, inner, base), _} == out_spec(plan, map, shape, fill, ops, 0n, 1n+inner, base) : Array<F32>}
289      {==}