May I ask if my description was not clear enough? Let me add some details: I extracted the model output CVPixelBuffer from the result returned by the MLModel.prediction() function and passed it as a cached outputBuffer to the outputBackings of MLPredictionOptions, like this: options.outputBackings = ["out": outputBuffer]. However, I still encountered the error: "Copy from tensor to pixel buffer (pixel_format_type: BGRA, image_pixel_type: BGR8, component_dtype: INT, component_pack: FMT_32) is not supported."