Pith. sign in

REVIEW 2 cited by

DBP: Discrimination Based Block-Level Pruning for Deep Model Acceleration

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1912.10178 v1 pith:RCSE3YGJ submitted 2019-12-21 cs.CV

DBP: Discrimination Based Block-Level Pruning for Deep Model Acceleration

classification cs.CV
keywords pruningratioaccelerationcomputationsfilter-levellayersmethodsparallel
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Neural network pruning is one of the most popular methods of accelerating the inference of deep convolutional neural networks (CNNs). The dominant pruning methods, filter-level pruning methods, evaluate their performance through the reduction ratio of computations and deem that a higher reduction ratio of computations is equivalent to a higher acceleration ratio in terms of inference time. However, we argue that they are not equivalent if parallel computing is considered. Given that filter-level pruning only prunes filters in layers and computations in a layer usually run in parallel, most computations reduced by filter-level pruning usually run in parallel with the un-reduced ones. Thus, the acceleration ratio of filter-level pruning is limited. To get a higher acceleration ratio, it is better to prune redundant layers because computations of different layers cannot run in parallel. In this paper, we propose our Discrimination based Block-level Pruning method (DBP). Specifically, DBP takes a sequence of consecutive layers (e.g., Conv-BN-ReLu) as a block and removes redundant blocks according to the discrimination of their output features. As a result, DBP achieves a considerable acceleration ratio by reducing the depth of CNNs. Extensive experiments show that DBP has surpassed state-of-the-art filter-level pruning methods in both accuracy and acceleration ratio. Our code will be made available soon.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Towards Efficient Convolutional Neural Network for Embedded Hardware via Multi-Dimensional Pruning

    cs.CV 2026-07 accept novelty 6.0

    TECO uses inner- and inter-dimensional importance scores plus a heuristic descent to prune CNN depth, width and resolution together, beating single-dimension SOTA on ImageNet accuracy and on-device latency.

  2. EdgeCompress: Coupling Multidimensional Model Compression and Dynamic Inference for EdgeAI

    cs.CV 2026-07 conditional novelty 4.0

    A three-stage CNN compression framework (dynamic cropping + compound shrinking + cascaded dynamic inference) reduces ResNet-50 MACs by 48.8% while improving ImageNet-1K top-1 accuracy by 0.8%.