# Post-training deep neural network pruning via layer-wise calibration

arXiv: https://arxiv.org/abs/2104.15023 · HTML: https://mlsanity.com/paper/2104.15023/

## Key facts

- Cited by 24 papers in the map (data through 14 Oct 2026).
- Builds on 18 of its 23 references that are in the map.
- First posted on arXiv 30 Apr 2021.

## Builds on (most cited first, 18 of 18)

- [Averaging Weights Leads to Wider Optima and Better Generalization](https://mlsanity.com/paper/1803.05407/) (2018)
- [Up or Down? Adaptive Rounding for Post-Training Quantization](https://mlsanity.com/paper/2004.10568/) (2020)
- [To prune, or not to prune: exploring the efficacy of pruning for model compression](https://mlsanity.com/paper/1710.01878/) (2017)
- [What is the State of Neural Network Pruning?](https://mlsanity.com/paper/2003.03033/) (2020)
- [BRECQ: Pushing the Limit of Post-Training Quantization by Block Reconstruction](https://mlsanity.com/paper/2102.05426/) (2021)
- [The State of Sparsity in Deep Neural Networks](https://mlsanity.com/paper/1902.09574/) (2019)
- [AMC: AutoML for Model Compression and Acceleration on Mobile Devices](https://mlsanity.com/paper/1802.03494/) (2018)
- [Data-Free Quantization Through Weight Equalization and Bias Correction](https://mlsanity.com/paper/1906.04721/) (2019)
- [Lookahead Optimizer: k steps forward, 1 step back](https://mlsanity.com/paper/1907.08610/) (2019)
- [Soft Threshold Weight Reparameterization for Learnable Sparsity](https://mlsanity.com/paper/2002.03231/) (2020)
- [Data-free parameter pruning for Deep Neural Networks](https://mlsanity.com/paper/1507.06149/) (2015)
- [Pre-training without Natural Images](https://mlsanity.com/paper/2101.08515/) (2021)
- [Improving Post Training Neural Quantization: Layer-wise Calibration and Integer Programming](https://mlsanity.com/paper/2006.10518/) (2020)
- [Fast Sparse ConvNets](https://mlsanity.com/paper/1911.09723/) (2019)
- [EagleEye: Fast Sub-net Evaluation for Efficient Neural Network Pruning](https://mlsanity.com/paper/2007.02491/) (2020)
- [Accelerating Sparse DNN Models without Hardware-Support via Tile-Wise Sparsity](https://mlsanity.com/paper/2008.13006/) (2020)
- [Can Vision Transformers Learn without Natural Images?](https://mlsanity.com/paper/2103.13023/) (2021)
- [Layer-Wise Data-Free CNN Compression](https://mlsanity.com/paper/2011.09058/) (2020)

## Built on by (most cited first, 24 of 24)

- [A Fast Post-Training Pruning Framework for Transformers](https://mlsanity.com/paper/2204.09656/) (2022)
- [ECoFLaP: Efficient Coarse-to-Fine Layer-Wise Pruning for Vision-Language Models](https://mlsanity.com/paper/2310.02998/) (2023)
- [Efficient Joint Optimization of Layer-Adaptive Weight Pruning in Deep Neural Networks](https://mlsanity.com/paper/2308.10438/) (2023)
- [Accurate Retraining-free Pruning for Pretrained Encoder-based Language Models](https://mlsanity.com/paper/2308.03449/) (2023)
- [Structurally Prune Anything: Any Architecture, Any Framework, Any Time](https://mlsanity.com/paper/2403.18955/) (2024)
- [EfficientLLaVA:Generalizable Auto-Pruning for Large Vision-language Models](https://mlsanity.com/paper/2503.15369/) (2025)
- [PTSBench: A Comprehensive Post-Training Sparsity Benchmark Towards Algorithms and Models](https://mlsanity.com/paper/2412.07268/) (2024)
- [Lossy and Lossless (L2) Post-training Model Size Compression](https://mlsanity.com/paper/2308.04269/) (2023)
- [Accelerating Learnt Video Codecs with Gradient Decay and Layer-wise Distillation](https://mlsanity.com/paper/2312.02605/) (2023)
- [SVD-NAS: Coupling Low-Rank Approximation and Neural Architecture Search](https://mlsanity.com/paper/2208.10404/) (2022)
- [QFT: Post-training quantization via fast joint finetuning of all degrees of freedom](https://mlsanity.com/paper/2212.02634/) (2022)
- [Generalizability of Mixture of Domain-Specific Adapters from the Lens of Signed Weight Directions and its Application to Effective Model Pruning](https://mlsanity.com/paper/2402.10639/) (2024)
- [Fast and Controllable Post-training Sparsity: Learning Optimal Sparsity Allocation with Global Constraint in Minutes](https://mlsanity.com/paper/2405.05808/) (2024)
- [FireFly-S: Exploiting Dual-Side Sparsity for Spiking Neural Networks Acceleration with Reconfigurable Spatial Architecture](https://mlsanity.com/paper/2408.15578/) (2024)
- [STAT: Shrinking Transformers After Training](https://mlsanity.com/paper/2406.00061/) (2024)
- [Spatio-Temporal Pruning for Compressed Spiking Large Language Models](https://mlsanity.com/paper/2508.20122/) (2025)
- [Onboard Optimization and Learning: A Survey](https://mlsanity.com/paper/2505.08793/) (2025)
- [Theoretical Compression Bounds for Wide Multilayer Perceptrons](https://mlsanity.com/paper/2512.06288/) (2025)
- [Adaptive Signal Resuscitation: Channel-wise Post-Pruning Repair for Sparse Vision Networks](https://mlsanity.com/paper/2605.21426/) (2026)
- [Relative Repairability: A Calibration-Based Diagnostic for High-Sparsity Post-Pruning Allocation](https://mlsanity.com/paper/2605.25508/) (2026)
- [Pruning On-the-Fly: A Recoverable Pruning Method without Fine-tuning](https://mlsanity.com/paper/2212.12651/) (2022)
- [Post-training Model Quantization Using GANs for Synthetic Data Generation](https://mlsanity.com/paper/2305.06052/) (2023)
- [TinySeg: Model Optimizing Framework for Image Segmentation on Tiny Embedded Systems](https://mlsanity.com/paper/2405.01857/) (2024)
- [How Sparsity Allocation Shapes Label-Free Post-Pruning Recoverability](https://mlsanity.com/paper/2605.21972/) (2026)

## Abstract (arXiv)

We present a post-training weight pruning method for deep neural networks that achieves accuracy levels tolerable for the production setting and that is sufficiently fast to be run on commodity hardware such as desktop CPUs or edge devices. We propose a data-free extension of the approach for computer vision models based on automatically-generated synthetic fractal images. We obtain state-of-the-art results for data-free neural network pruning, with ~1.5% top@1 accuracy drop for a ResNet50 on ImageNet at 50% sparsity rate. When using real data, we are able to get a ResNet50 model on ImageNet with 65% sparsity rate in 8-bit precision in a post-training setting with a ~1% top@1 accuracy drop. We release the code as a part of the OpenVINO(TM) Post-Training Optimization tool.