This also very much depends on the inference use case / context. For example, I work in deep learning on digital pathology where images can be up to 100000x1000

This also very much depends on the inference use case / context. For example, I work in deep learning on digital pathology where images can be up to 100000x100000pixels in size and inference needs GPUs as it's just way too slow otherwise.