Explain how a foundation model is built and used.
Separate large-scale learning from answering a user input.
Choose two correct statements.
Pretraining learns general patterns from large data, while inference computes an output from a trained model; ordinary inference does not update weights.
Detailed explanation
Inference and additional training are different.
It forms a model's base capabilities.
It produces a response at use time.
Bias and errors can affect capabilities and limits.
Check official documentation
Compare the following points with the official documentation.
AWS公式AIF-C01 Domain 2.1の基盤モデル、事前学習、推論を確認する。
Official references
What to confirm
学習時の処理と利用時の推論を区別して説明できる。
Key points
- Pretraining
- Inference
- Weights
Notes
- Environment: AWS公式AIF-C01試験ガイドとAWS公式ドキュメントの確認
- Compare the definition with the AIF-C01 objective and the service documentation before applying it.
- Keep the concept separate from a particular answer choice or implementation step.
Foundation review
Pretraining
Learning broad representations and patterns from a large dataset before task adaptation.
Inference
Computing a prediction or generated result by applying a trained model to an input.