An Empirical Study on Hyperparameter Optimization for Fine-Tuning Pre-trained Language Models
This addresses the problem of inefficient HPO for fine-tuning language models, which is incremental as it troubleshoots existing methods rather than introducing new ones.
The study investigated hyperparameter optimization (HPO) methods for fine-tuning pre-trained language models, finding that HPO often fails to outperform grid search due to insufficient time budget and overfitting, but can succeed with better settings, though overfitting persists in some cases.
The performance of fine-tuning pre-trained language models largely depends on the hyperparameter configuration. In this paper, we investigate the performance of modern hyperparameter optimization methods (HPO) on fine-tuning pre-trained language models. First, we study and report three HPO algorithms' performances on fine-tuning two state-of-the-art language models on the GLUE dataset. We find that using the same time budget, HPO often fails to outperform grid search due to two reasons: insufficient time budget and overfitting. We propose two general strategies and an experimental procedure to systematically troubleshoot HPO's failure cases. By applying the procedure, we observe that HPO can succeed with more appropriate settings in the search space and time budget; however, in certain cases overfitting remains. Finally, we make suggestions for future work. Our implementation can be found in https://github.com/microsoft/FLAML/tree/main/flaml/nlp/.