Learn Before
A team is developing a system to classify rare historical manuscripts. They have a small, highly specialized dataset. One engineer argues against using a large, generally pre-trained model, stating, 'Our task is too unique. A model trained on the entire internet will have learned irrelevant patterns. We should build a smaller model from scratch using only our specific manuscript data.' Which of the following statements best evaluates this engineer's argument?
0
1
Tags
Ch.1 Pre-training - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Evaluation in Bloom's Taxonomy
Cognitive Psychology
Psychology
Social Science
Empirical Science
Science
Related
A team is developing a system to classify rare historical manuscripts. They have a small, highly specialized dataset. One engineer argues against using a large, generally pre-trained model, stating, 'Our task is too unique. A model trained on the entire internet will have learned irrelevant patterns. We should build a smaller model from scratch using only our specific manuscript data.' Which of the following statements best evaluates this engineer's argument?
Model Development Strategy for NLP Products
The Generalist Advantage in Model Pre-training