試験
通常
試験時間: 00:00問題時間: 00:00
問題一覧
問題 57 / 65
Which method is most effective when you want to efficiently adapt to many downstream tasks while minimizing computational resource and storage costs?
①Fine-tune all model parameters.
②Parameter-efficient methods such as LoRA and adapters
③Perform large-scale updates through continual pretraining.
④Task-specific prompt tuning
⑤Create a small model using model distillation.