試験
通常
試験時間: 00:00問題時間: 00:00
問題一覧
問題 27 / 65
I want to reduce energy consumption during inference with little or no additional retraining. Which technique is most appropriate?
①Increase the size of the model ensemble to stabilize.
②Reduce the learning rate and retrain.
③Quantize the model and perform inference at low precision.
④Perform inference at full precision every time.
⑤Preprocess all input data in the cloud.
