Model Fusion through Bayesian Optimization in Language Model Fine-Tuning
This addresses the engineering challenges in fine-tuning for practitioners, though it is incremental as it builds on existing model fusion and Bayesian optimization methods.
The paper tackles the problem of selecting the best model during fine-tuning of pre-trained language models by introducing a model fusion technique that optimizes both metric and loss through multi-objective Bayesian optimization, resulting in considerable performance improvements across various downstream tasks.
Fine-tuning pre-trained models for downstream tasks is a widely adopted technique known for its adaptability and reliability across various domains. Despite its conceptual simplicity, fine-tuning entails several troublesome engineering choices, such as selecting hyperparameters and determining checkpoints from an optimization trajectory. To tackle the difficulty of choosing the best model, one effective solution is model fusion, which combines multiple models in a parameter space. However, we observe a large discrepancy between loss and metric landscapes during the fine-tuning of pre-trained language models. Building on this observation, we introduce a novel model fusion technique that optimizes both the desired metric and loss through multi-objective Bayesian optimization. In addition, to effectively select hyperparameters, we establish a two-stage procedure by integrating Bayesian optimization processes into our framework. Experiments across various downstream tasks show considerable performance improvements using our Bayesian optimization-guided method.