4 ms·When the inference is done on the device otherwise it's just an AI thin client.by Zetobal 3y agoWhen the inference is done on the device otherwise it's just an AI thin client.