On the Predictability of Pruning Across Scales
arXiv:2006.10621
Abstract
We show that the error of iteratively magnitude-pruned networks empirically follows a scaling law with interpretable coefficients that depend on the architecture and task. We functionally approximate the error of the pruned networks, showing it is predictable in terms of an invariant tying width, depth, and pruning level, such that networks of vastly different pruned densities are interchangeable. We demonstrate the accuracy of this approximation over orders of magnitude in depth, width, dataset size, and density. We show that the functional form holds (generalizes) for large scale data (e.g., ImageNet) and architectures (e.g., ResNets). As neural networks become ever larger and costlier to train, our findings suggest a framework for reasoning conceptually and analytically about a standard method for unstructured pruning.
References in corpus (6)
- EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
- Language Models are Few-Shot Learners
- Scaling Laws for Neural Language Models
- The State of Sparsity in Deep Neural Networks
- Deep Learning Scaling is Predictable, Empirically
- Comparing Rewinding and Fine-tuning in Neural Network Pruning