def forward()

in neuralcompression/models/deep_video_compression.py [0:0]

• 10 lines of code
• 4 McCabe index (conditional complexity)

    def forward(self, image: Tensor) -> Tensor:
        upsample_size = tuple(image.shape[2:])

        # residual block includes skipped connection
        image = self.input_block(image)

        # if we have a child, downsample, run it, and then upsample back
        if self.child is not None:
            image = image + F.interpolate(
                self.child(F.avg_pool2d(image, kernel_size=2)),
                size=upsample_size,
                mode="nearest",
            )

        # residual block includes skipped connection
        return self.output_block(image)