Exploring the Performance of Pruning Methods in Neural Networks: An Empirical Study of the Lottery Ticket Hypothesis
This is an incremental study for researchers interested in neural network compression, focusing on benchmarking pruning techniques without major breakthroughs.
The paper empirically compares pruning methods like L1 unstructured, Fisher, and random pruning in neural networks, testing scenarios such as one-shot vs. iterative pruning and varying network widths, and proposes batched Fisher pruning for efficiency.
In this paper, we explore the performance of different pruning methods in the context of the lottery ticket hypothesis. We compare the performance of L1 unstructured pruning, Fisher pruning, and random pruning on different network architectures and pruning scenarios. The experiments include an evaluation of one-shot and iterative pruning, an examination of weight movement in the network during pruning, a comparison of the pruning methods on networks of varying widths, and an analysis of the performance of the methods when the network becomes very sparse. Additionally, we propose and evaluate a new method for efficient computation of Fisher pruning, known as batched Fisher pruning.