試験
通常
試験時間: 00:00問題時間: 00:00
問題一覧
問題 42 / 65
What is the most effective option for processing a large number of batch generation jobs while reducing inference costs?
①Always run the latest and largest model on a single node.
②Perform full inference every time in real time.
③Process only on edge devices.
④Run using a combination of batch processing and spot instances.
⑤Do nothing but strictly shorten the prompt length.