공식 발표해외
Mix, Don't Tune: Bilingual Pre-Training Outperforms Hyperparameter Search in Data-Constrained Settings
Apple ML Research
For most languages of the world, language model pre-training operates in a data-constrained regime where models must repeat their training data many times, degrading generalization.
원문의 저작권은 Apple ML Research에 있습니다. 위 발췌는 원문 확인을 돕기 위한 짧은 인용입니다.