How to Tune a Multilingual Encoder Model for Germanic Languages: A Study of PEFT, Full Fine-Tuning, and Language Adapters
This work addresses the problem of optimizing fine-tuning strategies for multilingual models in specific Germanic languages, but it is incremental as it builds on existing PEFT and adapter research without introducing new methods.
This paper investigated how to best fine-tune the multilingual encoder model mDeBERTa for tasks in German, Swedish, and Icelandic, finding that parameter-efficient fine-tuning (PEFT) methods like LoRA and adapters were more effective for higher-resource German, but results varied for Swedish and Icelandic, with PEFT better for question answering and full fine-tuning preferable for named entity recognition.
This paper investigates the optimal use of the multilingual encoder model mDeBERTa for tasks in three Germanic languages -- German, Swedish, and Icelandic -- representing varying levels of presence and likely data quality in mDeBERTas pre-training data. We compare full fine-tuning with the parameter-efficient fine-tuning (PEFT) methods LoRA and Pfeiffer bottleneck adapters, finding that PEFT is more effective for the higher-resource language, German. However, results for Swedish and Icelandic are less consistent. We also observe differences between tasks: While PEFT tends to work better for question answering, full fine-tuning is preferable for named entity recognition. Inspired by previous research on modular approaches that combine task and language adapters, we evaluate the impact of adding PEFT modules trained on unstructured text, finding that this approach is not beneficial.