Transfer Learning for Structured Pruning under Limited Task Data
This work addresses a bottleneck for deploying large models in resource-constrained applications by reducing data requirements, though it is incremental as it builds on existing pruning and transfer learning methods.
The paper tackles the problem of structured pruning requiring large amounts of task-specific data by proposing a framework that combines pruning with transfer learning, resulting in pruned models with improved generalization over baselines.
Large, pre-trained models are problematic to use in resource constrained applications. Fortunately, task-aware structured pruning methods offer a solution. These approaches reduce model size by dropping structural units like layers and attention heads in a manner that takes into account the end-task. However, these pruning algorithms require more task-specific data than is typically available. We propose a framework which combines structured pruning with transfer learning to reduce the need for task-specific data. Our empirical results answer questions such as: How should the two tasks be coupled? What parameters should be transferred? And, when during training should transfer learning be introduced? Leveraging these insights, we demonstrate that our framework results in pruned models with improved generalization over strong baselines.