Compacting an AI model to run faster. AI quantization is primarily performed at the inference side (user side) so that it can run more quickly in phones and desktop computers. For example, whereas the ...