BUG: VkFFTConfiguration::bufferSize, inputBufferSize, and outputBufferSize should be in bytes

[Leengit]

They currently have the number of entries in the image, but need to be further multiplied by the size of the pixel types in bytes.