# Prune-Quantize-Distill: An Ordered Pipeline for Efficient Neural Network Compression

arXiv: https://arxiv.org/abs/2604.04988 · HTML: https://mlsanity.com/paper/2604.04988/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 12 of its 16 references that are in the map.
- First posted on arXiv 5 Apr 2026.

## Builds on (most cited first, 12 of 12)

- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks](https://mlsanity.com/paper/1803.03635/) (2018)
- [SNIP: Single-shot Network Pruning based on Connection Sensitivity](https://mlsanity.com/paper/1810.02340/) (2018)
- [Once-for-All: Train One Network and Specialize it for Efficient Deployment](https://mlsanity.com/paper/1908.09791/) (2019)
- [Contrastive Representation Distillation](https://mlsanity.com/paper/1910.10699/) (2019)
- [Learned Step Size Quantization](https://mlsanity.com/paper/1902.08153/) (2019)
- [DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients](https://mlsanity.com/paper/1606.06160/) (2016)
- [Pruning and Quantization for Deep Neural Network Acceleration: A Survey](https://mlsanity.com/paper/2101.09671/) (2021)
- [Incremental Network Quantization: Towards Lossless CNNs with Low-Precision Weights](https://mlsanity.com/paper/1702.03044/) (2017)
- [A Survey on Deep Neural Network Pruning-Taxonomy, Comparison, Analysis, and Recommendations](https://mlsanity.com/paper/2308.06767/) (2023)
- [PQK: Model Compression via Pruning, Quantization, and Knowledge Distillation](https://mlsanity.com/paper/2106.14681/) (2021)
- [Boosting Pruned Networks with Linear Over-parameterization](https://mlsanity.com/paper/2204.11444/) (2022)

## Built on by (most cited first, 1 of 1)

- [LightVesselNet: An Ultra-Lightweight Sub-100K Parameter Network for Retinal Blood Vessel Segmentation](https://mlsanity.com/paper/2606.05354/) (2026)

## Abstract (arXiv)

Modern deployment often requires trading accuracy for efficiency under tight CPU and memory constraints, yet common compression proxies such as parameter count or FLOPs do not reliably predict wall-clock inference time. In particular, unstructured sparsity can reduce model storage while failing to accelerate (and sometimes slightly slowing down) standard CPU execution due to irregular memory access and sparse kernel overhead. Motivated by this gap between compression and acceleration, we study a practical, ordered pipeline that targets measured latency by combining three widely used techniques: unstructured pruning, INT8 quantization-aware training (QAT), and knowledge distillation (KD). Empirically, INT8 QAT provides the dominant runtime benefit, while pruning mainly acts as a capacity-reduction pre-conditioner that improves the robustness of subsequent low-precision optimization; KD, applied last, recovers accuracy within the already constrained sparse INT8 regime without changing the deployment form. We evaluate on CIFAR-10/100 using three backbones (ResNet-18, WRN-28-10, and VGG-16-BN). Across all settings, the ordered pipeline achieves a stronger accuracy-size-latency frontier than any single technique alone, reaching 0.99-1.42 ms CPU latency with competitive accuracy and compact checkpoints. Controlled ordering ablations with a fixed 20/40/40 epoch allocation further confirm that stage order is consequential, with the proposed ordering generally performing best among the tested permutations. Overall, our results provide a simple guideline for edge deployment: evaluate compression choices in the joint accuracy-size-latency space using measured runtime, rather than proxy metrics alone.