Overcoming Vocabulary Mismatch: Vocabulary-agnostic Teacher Guided Language Modeling
This addresses vocabulary mismatch in knowledge distillation for language modeling, offering a robust solution for efficient training with mismatched vocabularies.
The paper tackles the problem of vocabulary mismatches between teacher and student language models, which cause divergent token sequences and output distributions, by proposing Vocabulary-agnostic Teacher Guided Language Modeling (VocAgnoLM) with token-level lexical alignment and teacher guided loss. It achieves a 46% performance improvement with a 1B student model using a teacher sharing only 6% vocabulary.
Using large teacher models to guide the training of smaller student models has become the prevailing paradigm for efficient and effective learning. However, vocabulary mismatches between teacher and student language models pose significant challenges in language modeling, resulting in divergent token sequences and output distributions. To overcome these limitations, we propose Vocabulary-agnostic Teacher Guided Language Modeling (VocAgnoLM), a novel approach that bridges the gap caused by vocabulary mismatch through two key methods: (1) Token-level Lexical Alignment, which aligns token sequences across mismatched vocabularies, and (2) Teacher Guided Loss, which leverages the loss of teacher model to guide effective student training. We demonstrate its effectiveness in language modeling with 1B student model using various 7B teacher models with different vocabularies. Notably, with Qwen2.5-Math-Instruct, a teacher model sharing only about 6% of its vocabulary with TinyLlama, VocAgnoLM achieves a 46% performance improvement compared to naive continual pretraining. Furthermore, we demonstrate that VocAgnoLM consistently benefits from stronger teacher models, providing a robust solution to vocabulary mismatches in language modeling.