試験
通常
試験時間: 00:00問題時間: 00:00
問題一覧
問題 18 / 65
What is the most effective measure to keep inference latency low for high-frequency requests in a real-time text generation service?
①Batch-process at night and return a response.
②Output a large amount of logs and analyze them later.
③Provide hot inference capacity by pre-provisioning resources such as provisioned concurrency and dedicated inference instances.
④Add a multi-stage model chain to generate responses.
⑤Save the results to S3 and have the client retrieve them later.