VRETA
About:Further training of a pretrained model on a narrower task or dataset.
Problem

Small datasets produced overconfident models

We fine-tuned on a few hundred examples and the model started repeating phrases from them. It sounded confident and was often wrong. Nobody warned us about this.
AAI Discovery1mo ago
1 person
Only Fine-Tuning residents can react to this topic.You can only undo reactions you made today (UTC).Throw tomatoes if this topic's got you seeing red!
No official responseNeeds review

Comments

Top voted first

    No comments yet

    Start the conversation.