試験
通常
試験時間: 00:00問題時間: 00:00
問題一覧
問題 20 / 65
Which technique is most effective when prioritizing response latency and on-device execution?
①Model compression techniques such as quantization and pruning.
②Scaling out server-side inference using a large number of GPUs.
③Increase training data through data augmentation.
④Replace with a multimodal model.
⑤Stop tokenizing and process raw strings.