フリー問題

問題文

A pipeline holds a tokenizer and a hashing step and a classifier and the team calls the fit method of the pipeline on the training frame. What comes back from that call?

選択肢

1. A fitted pipeline comes back and every learning step in it has turned into a step that holds a transform of its own.
2. A frame comes back and the pipeline discards the fitted steps that the last stage of the run produced.
3. A fitted pipeline comes back and every learning step inside it stays an estimator that fits itself again in the run at the moment of prediction.
4. A metric comes back and the pipeline keeps the fitted steps that an internal cache holds for the next call.