mlsanityData to 7 Oct 2026

Speedup deep learning models on GPU by taking advantage of efficient unstructured pruning and bit-width reduction

arXiv 2112.15445 ↗First posted 28 Dec 2021cs.LG cs.AI cs.PFcited by 3

Not on the map. Why?

Abstract (arXiv)

This work is focused on the pruning of some convolutional neural networks (CNNs) and improving theirs efficiency on graphic processing units (GPU) by using a direct sparse algorithm. The Nvidia deep neural network (cuDnn) library is the most effective implementations of deep learning (DL) algorithms for GPUs. GPUs are the most commonly used accelerators for deep learning computations. One of the most common techniques for improving the efficiency of CNN models is weight pruning and quantization. There are two main types of pruning: structural and non-structural. The first enables much easier acceleration on many type of accelerators, but with this type it is difficult to achieve a sparsity level and accuracy as high as that obtained with the second type. Non-structural pruning with retraining can generate a weight tensors up to 90% or more of sparsity in some deep CNN models. In this article the pruning algorithm is presented which makes it possible to achieve high sparsity levels without accuracy drop. In the next stage the linear and non-linear quantization is adapted for further time and footprint reduction. This paper is an extended of previously published paper concerning effective pruning techniques and present real models pruned with high sparsities and reduced precision which can achieve better performance than the CuDnn library.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 29 of its 44 references that are in the map. First posted on arXiv 28 Dec 2021.

Builds on

29 of 44 references.

References · 29

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2015Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networkscited by 9,617
  3. 2014Going Deeper with Convolutionscited by 4,231
  4. 2015Learning both Weights and Connections for Efficient Neural Networkscited by 1,998
  5. 2016Pruning Filters for Efficient ConvNetscited by 1,129
  6. 2014Convolutional Neural Networks for Sentence Classificationcited by 1,124
  7. 2018Detecting Spacecraft Anomalies Using LSTMs and Nonparametric Dynamic Thresholdingcited by 407
  8. 2020Movement Pruning: Adaptive Sparsity by Fine-Tuningcited by 398
  9. 2016EIE: Efficient Inference Engine on Compressed Deep Neural Networkcited by 291
  10. 2017Variational Dropout Sparsifies Deep Neural Networkscited by 265
  11. 2014cuDNN: Efficient Primitives for Deep Learningcited by 240
  12. 2018Discrimination-aware Channel Pruning for Deep Neural Networkscited by 171
17 more
  1. 2015Fixed Point Quantization of Deep Convolutional Networkscited by 141
  2. 2015Fast Algorithms for Convolutional Neural Networkscited by 102
  3. 2017Comparative Study of CNN and RNN for Natural Language Processingcited by 81
  4. 2016Going Deeper with Contextual CNN for Hyperspectral Image Classificationcited by 58
  5. 2018ADMM-NN: An Algorithm-Hardware Co-Design Framework of DNNs Using Alternating Direction Method of Multiplierscited by 42
  6. 2016Faster CNNs with Direct Sparse Convolutions and Guided Pruningcited by 39
  7. 2017Structured Bayesian Pruning via Log-Normal Multiplicative Noisecited by 39
  8. 2019Stabilizing the Lottery Ticket Hypothesiscited by 36
  9. 2018Learning to Prune Filters in Convolutional Neural Networkscited by 25
  10. 2018Escoin: Efficient Sparse Convolutional Neural Network Inference on GPUscited by 11
  11. 2018A novel channel pruning method for deep neural network compressioncited by 8
  12. 2021Ensemble neuroevolution based approach for multivariate time series anomaly detectioncited by 8
  13. 2016Ristretto: Hardware-Oriented Approximation of Convolutional Neural Networkscited by 6
  14. 2020Retrain or not retrain? -- efficient pruning methods of deep CNN networkscited by 6
  15. 2020Gradual Channel Pruning while Training using Feature Relevance Scores for Convolutional Neural Networkscited by 6
  16. 2018Reliable Identification of Redundant Kernels for Convolutional Neural Network Compressioncited by 2
  17. 2019GPU Fast Convolution via the Overlap-and-Save Method in Shared Memorycited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 1; 2025: 0; 2026: 2.

3 papers.

  1. 2024Towards efficient deep autoencoders for multivariate time series anomaly detectioncited by 0
  2. 2026Evolutionary fine tuning of quantized convolution-based deep learning modelscited by 0
  3. 2026Rethinking Depth Pruning for Vision Transformers: A Heterogeneity-Aware Perspectivecited by 0