[relay][pass] add layerNormal infer layout

[chengven027]

The structure before adding is

fn (%x: Tensor[(1, 56, 56, 64), float32] /* ty=Tensor[(1, 56, 56, 64), float32] */, %weight: Tensor[(3, 3, 64, 64), float32] /* ty=Tensor[(3, 3, 64, 64), float32] */, %gamma: Tensor[(64), float32] /* ty=Tensor[(64), float32] */, %beta: Tensor[(64), float32] /* ty=Tensor[(64), float32] */) -> Tensor[(1, 56, 56, 64), float32] {
  %0 = layout_transform(%x, src_layout="NHWC", dst_layout="NCHW") /* ty=Tensor[(1, 64, 56, 56), float32] */;
  %1 = layout_transform(%weight, src_layout="HWIO", dst_layout="OIHW") /* ty=Tensor[(64, 64, 3, 3), float32] */;
  %2 = nn.conv2d(%0, %1, padding=[1, 1, 1, 1], channels=64, kernel_size=[3, 3]) /* ty=Tensor[(1, 64, 56, 56), float32] */;
  %3 = layout_transform(%2, src_layout="NCHW", dst_layout="NHWC") /* ty=Tensor[(1, 56, 56, 64), float32] */;
  nn.layer_norm(%3, %gamma, %beta, axis=3) /* ty=Tensor[(1, 56, 56, 64), float32] */
} /* ty=fn (Tensor[(1, 56, 56, 64), float32], Tensor[(3, 3, 64, 64), float32], Tensor[(64), float32], Tensor[(64), float32]) -> Tensor[(1, 56, 56, 64), float32] */

The structure after adding is

fn (%x: Tensor[(1, 56, 56, 64), float32] /* ty=Tensor[(1, 56, 56, 64), float32] */, %weight: Tensor[(3, 3, 64, 64), float32] /* ty=Tensor[(3, 3, 64, 64), float32] */, %gamma: Tensor[(64), float32] /* ty=Tensor[(64), float32] */, %beta: Tensor[(64), float32] /* ty=Tensor[(64), float32] */) -> Tensor[(1, 56, 56, 64), float32] {
  %0 = layout_transform(%x, src_layout="NHWC", dst_layout="NCHW") /* ty=Tensor[(1, 64, 56, 56), float32] */;
  %1 = layout_transform(%weight, src_layout="HWIO", dst_layout="OIHW") /* ty=Tensor[(64, 64, 3, 3), float32] */;
  %2 = nn.conv2d(%0, %1, padding=[1, 1, 1, 1], channels=64, kernel_size=[3, 3]) /* ty=Tensor[(1, 64, 56, 56), float32] */;
  %3 = nn.layer_norm(%2, %gamma, %beta, axis=1) /* ty=Tensor[(1, 64, 56, 56), float32] */;
  layout_transform(%3, src_layout="NCHW", dst_layout="NHWC") /* ty=Tensor[(1, 56, 56, 64), float32] */
} /* ty=fn (Tensor[(1, 56, 56, 64), float32], Tensor[(3, 3, 64, 64), float32], Tensor[(64), float32], Tensor[(64), float32]) -> Tensor[(1, 56, 56, 64), float32] */