Movement Pruning: Adaptive Sparsity by Fine-Tuning

Code

• github.com/huggingface/transformers/tree/master/examples/movement-pruning
  Officialpytorch★ 0
• github.com/huggingface/nn_pruning
  pytorch★ 406
• github.com/huggingface/block_movement_pruning
  pytorch★ 83
• github.com/kainoj/pruning-bias
  pytorch★ 7

Abstract

Magnitude pruning is a widely used strategy for reducing model size in pure supervised learning; however, it is less effective in the transfer learning regime that has become standard for state-of-the-art natural language processing applications. We propose the use of movement pruning, a simple, deterministic first-order weight pruning method that is more adaptive to pretrained model fine-tuning. We give mathematical foundations to the method and compare it to existing zeroth- and first-order pruning methods. Experiments show that when pruning large pretrained language models, movement pruning shows significant improvements in high-sparsity regimes. When combined with distillation, the approach achieves minimal accuracy loss with down to only 3% of the model parameters.

Tasks

Reproductions