S2HPruner: Soft-to-Hard Distillation Bridges the Discretization Gap in Pruning
Weihao Lin, Shengji Tang, Chong Yu, Peng Ye, Tao Chen

TL;DR
S2HPruner introduces a one-stage distillation framework that effectively bridges the discretization gap in network pruning, leading to superior performance without fine-tuning across multiple benchmarks.
Contribution
It proposes a novel structural differentiable mask pruning method with bidirectional knowledge distillation to address the discretization gap in pruning.
Findings
Achieves better pruning performance without fine-tuning.
Effective across various datasets and network architectures.
Outperforms existing pruning methods in benchmarks.
Abstract
Recently, differentiable mask pruning methods optimize the continuous relaxation architecture (soft network) as the proxy of the pruned discrete network (hard network) for superior sub-architecture search. However, due to the agnostic impact of the discretization process, the hard network struggles with the equivalent representational capacity as the soft network, namely discretization gap, which severely spoils the pruning performance. In this paper, we first investigate the discretization gap and propose a novel structural differentiable mask pruning framework named S2HPruner to bridge the discretization gap in a one-stage manner. In the training procedure, SH2Pruner forwards both the soft network and its corresponding hard network, then distills the hard network under the supervision of the soft network. To optimize the mask and prevent performance degradation, we propose a decoupled…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsInnovations in Concrete and Construction Materials · Membrane Separation Technologies
MethodsPruning
