Function upfirdn2d_native

src/diffusers/models/upsampling.py:421–468

(
    tensor: torch.Tensor,
    kernel: torch.Tensor,
    up: int = 1,
    down: int = 1,
    pad: tuple[int, int] = (0, 0),
)

def upfirdn2d_native(
    tensor: torch.Tensor,
    kernel: torch.Tensor,
    up: int = 1,
    down: int = 1,
    pad: tuple[int, int] = (0, 0),
) -> torch.Tensor:
    up_x = up_y = up
    down_x = down_y = down
    pad_x0 = pad_y0 = pad[0]
    pad_x1 = pad_y1 = pad[1]

    _, channel, in_h, in_w = tensor.shape
    tensor = tensor.reshape(-1, in_h, in_w, 1)

    _, in_h, in_w, minor = tensor.shape
    kernel_h, kernel_w = kernel.shape

    out = tensor.view(-1, in_h, 1, in_w, 1, minor)
    out = F.pad(out, [0, 0, 0, up_x - 1, 0, 0, 0, up_y - 1])
    out = out.view(-1, in_h * up_y, in_w * up_x, minor)

    out = F.pad(out, [0, 0, max(pad_x0, 0), max(pad_x1, 0), max(pad_y0, 0), max(pad_y1, 0)])
    out = out.to(tensor.device)  # Move back to mps if necessary
    out = out[
        :,
        max(-pad_y0, 0) : out.shape[1] - max(-pad_y1, 0),
        max(-pad_x0, 0) : out.shape[2] - max(-pad_x1, 0),
        :,
    ]

    out = out.permute(0, 3, 1, 2)
    out = out.reshape([-1, 1, in_h * up_y + pad_y0 + pad_y1, in_w * up_x + pad_x0 + pad_x1])
    w = torch.flip(kernel, [0, 1]).view(1, 1, kernel_h, kernel_w)
    out = F.conv2d(out, w)
    out = out.reshape(
        -1,
        minor,
        in_h * up_y + pad_y0 + pad_y1 - kernel_h + 1,
        in_w * up_x + pad_x0 + pad_x1 - kernel_w + 1,
    )
    out = out.permute(0, 2, 3, 1)
    out = out[:, ::down_y, ::down_x, :]

    out_h = (in_h * up_y + pad_y0 + pad_y1 - kernel_h) // down_y + 1
    out_w = (in_w * up_x + pad_x0 + pad_x1 - kernel_w) // down_x + 1

    return out.view(-1, channel, out_h, out_w)


def upsample_2d(