mlsanityData to 7 Oct 2026

Mixed-Precision Neural Network Quantization via Learned Layer-wise Importance

arXiv 2203.08368 ↗First posted 16 Mar 2022cs.LG cs.CVcited by 53

Not on the map. Why?

Abstract (arXiv)

The exponentially large discrete search space in mixed-precision quantization (MPQ) makes it hard to determine the optimal bit-width for each layer. Previous works usually resort to iterative search methods on the training set, which consume hundreds or even thousands of GPU-hours. In this study, we reveal that some unique learnable parameters in quantization, namely the scale factors in the quantizer, can serve as importance indicators of a layer, reflecting the contribution of that layer to the final accuracy at certain bit-widths. These importance indicators naturally perceive the numerical transformation during quantization-aware training, which can precisely provide quantization sensitivity metrics of layers. However, a deep network always contains hundreds of such indicators, and training them one by one would lead to an excessive time cost. To overcome this issue, we propose a joint training scheme that can obtain all indicators at once. It considerably speeds up the indicators training process by parallelizing the original sequential training processes. With these learned importance indicators, we formulate the MPQ search problem as a one-time integer linear programming (ILP) problem. That avoids the iterative search and significantly reduces search time without limiting the bit-width search space. For example, MPQ search on ResNet18 with our indicators takes only 0.06 s, which improves time efficiency exponentially compared to iterative search methods. Also, extensive experiments show our approach can achieve SOTA accuracy on ImageNet for far-ranging models with various constraints (e.g., BitOps, compress rate). Code is available on https://github.com/1hunters/LIMPQ.

Key facts. Cited by 53 papers in the map (data through 7 Oct 2026). Builds on 26 of its 30 references that are in the map. First posted on arXiv 16 Mar 2022.

Builds on

26 of 30 references.

References · 26

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,827
  3. 2017Learning Efficient Convolutional Networks through Network Slimmingcited by 673
  4. 2019Learned Step Size Quantizationcited by 604
  5. 2016DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradientscited by 557
  6. 2018PACT: Parameterized Clipping Activation for Quantized Neural Networkscited by 456
  7. 2018HAQ: Hardware-Aware Automated Quantization with Mixed Precisioncited by 426
  8. 2019Single Path One-Shot Neural Architecture Search with Uniform Samplingcited by 411
  9. 2019HAWQ: Hessian AWare Quantization of Neural Networks with Mixed-Precisioncited by 387
  10. 2018LQ-Nets: Learned Quantization for Highly Accurate and Compact Deep Neural Networkscited by 272
  11. 2019HAWQ-V2: Hessian Aware trace-Weighted Quantization of Neural Networkscited by 250
  12. 2017Incremental Network Quantization: Towards Lossless CNNs with Low-Precision Weightscited by 211
14 more
  1. 2020LSQ+: Improving low-bit quantization through learnable offsets and better initializationcited by 180
  2. 2018Learning to Quantize Deep Networks by Optimizing Quantization Intervals with Task Losscited by 157
  3. 2019FairNAS: Rethinking Evaluation Fairness of Weight Sharing Neural Architecture Searchcited by 129
  4. 2018Mixed Precision Quantization of ConvNets via Differentiable Neural Architecture Searchcited by 122
  5. 2017Deep Learning with Low Precision by Half-wave Gaussian Quantizationcited by 115
  6. 2020Rethinking Differentiable Search for Mixed-Precision Neural Networkscited by 79
  7. 2019AutoQ: Automated Kernel-Wise Neural Network Quantizationcited by 60
  8. 2021Towards Mixed-Precision Quantization of Neural Networks via Constrained Optimizationcited by 57
  9. 2020PROFIT: A Novel Training Method for sub-4-bit MobileNet Modelscited by 50
  10. 2021Differentiable Model Compression via Pseudo Quantization Noisecited by 44
  11. 2020HMQ: Hardware Friendly Mixed Precision Quantization Block for CNNscited by 43
  12. 2020Search What You Want: Barrier Panelty NAS for Mixed Precision Quantizationcited by 31
  13. 2018NICE: Noise Injection and Clamping Estimation for Neural Network Quantizationcited by 24
  14. 2021BN-NAS: Neural Architecture Search with Batch Normalizationcited by 21

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 3; 2023: 12; 2024: 17; 2025: 15; 2026: 6.

53 papers.

Sort:

By area:

  1. 2023Outlier Weighed Layerwise Sparsity (OWL): A Missing Secret Sauce for Pruning LLMs to High Sparsitycited by 163
  2. 2023LQ-LoRA: Low-rank Plus Quantized Matrix Decomposition for Efficient Language Model Finetuningcited by 92
  3. 2024Q-DiT: Accurate Post-Training Quantization for Diffusion Transformerscited by 91
  4. 2024NetLLM: Adapting Large Language Models for Networkingcited by 76
  5. 2022A Comprehensive Survey on Model Quantization for Deep Neural Networks in Image Classificationcited by 64
  6. 2025SP-VLA: A Joint Model Scheduling and Token Pruning Approach for VLA Model Accelerationcited by 56
  7. 2024MesonGS: Post-training Compression of 3D Gaussians via Efficient Attribute Transformationcited by 54
  8. 2024AlphaPruning: Using Heavy-Tailed Self Regularization Theory for Improved Layer-wise Pruning of Large Language Modelscited by 35
  9. 2023A Machine Learning-oriented Survey on Tiny Machine Learningcited by 26
  10. 2023Data Quality-aware Mixed-precision Quantization via Hybrid Reinforcement Learningcited by 17
  11. 2024LLMEasyQuant: Scalable Quantization for Parallel and Distributed LLM Inferencecited by 15
  12. 2024Retraining-free Model Quantization via One-Shot Weight-Coupling Learningcited by 13
38 more
  1. 2023Mixed-Precision Quantization for Federated Learning on Resource-Constrained Heterogeneous Devicescited by 11
  2. 2025Low-bit Model Quantization for Deep Neural Networks: A Surveycited by 11
  3. 2022FIT: A Metric for Model Sensitivitycited by 10
  4. 2024PRANCE: Joint Token-Optimization and Structural Channel-Pruning for Adaptive ViT Inferencecited by 10
  5. 2024Investigating the Impact of Quantization on Adversarial Robustnesscited by 9
  6. 2024TMPQ-DM: Joint Timestep Reduction and Quantization Precision Selection for Efficient Diffusion Modelscited by 9
  7. 2025Mix-QViT: Mixed-Precision Vision Transformer Quantization Driven by Layer Importance and Quantization Sensitivitycited by 9
  8. 2025Win Fast or Lose Slow: Balancing Speed and Accuracy in Latency-Sensitive Decisions of LLMscited by 9
  9. 2024AdaBM: On-the-Fly Adaptive Bit Mapping for Image Super-Resolutioncited by 8
  10. 2023SEAM: Searching Transferable Mixed-Precision Quantization Policy through Large Margin Regularizationcited by 6
  11. 2024TransAxx: Efficient Transformers with Approximate Computingcited by 6
  12. 2024Content-Aware Radiance Fields: Aligning Model Complexity with Scene Intricacy Through Learned Bitwidth Quantizationcited by 6
  13. 2023PreQuant: A Task-agnostic Quantization Approach for Pre-trained Language Modelscited by 5
  14. 2023Data-Aware Gradient Compression for FL in Communication-Constrained Mobile Computingcited by 5
  15. 2025Mix-QSAM: Mixed-Precision Quantization of the Segment Anything Modelcited by 4
  16. 2025InfoQ: Mixed-Precision Quantization via Global Information Flowcited by 4
  17. 2022AMED: Automatic Mixed-Precision Quantization for Edge Devicescited by 3
  18. 2023Towards Optimal Compression: Joint Pruning and Quantizationcited by 3
  19. 2023Sensitivity-Aware Mixed-Precision Quantization and Width Optimization of Deep Neural Networks Through Cluster-Based Tree-Structured Parzen Estimationcited by 3
  20. 2025A General Error-Theoretical Analysis Framework for Constructing Compression Strategiescited by 3
  21. 2024Real-Time Spacecraft Pose Estimation Using Mixed-Precision Quantized Neural Network on COTS Reconfigurable MPSoCcited by 2
  22. 2024TaQ-DiT: Time-aware Quantization for Diffusion Transformerscited by 2
  23. 2024GAQAT: gradient-adaptive quantization-aware training for domain generalizationcited by 2
  24. 2025Quantization Meets OOD: Generalizable Quantization-aware Training from a Flatness Perspectivecited by 2
  25. 2025VVS: Accelerating Speculative Decoding for Visual Autoregressive Generation via Partial Verification Skippingcited by 2
  26. 2026Tail-Aware Post-Training Quantization for 3D Geometry Modelscited by 2
  27. 2024ARQ: A Mixed-Precision Quantization Framework for Accurate and Certifiably Robust DNNscited by 1
  28. 2025Coded Deep Learning: Framework and Algorithmcited by 1
  29. 2025Regularization-based Framework for Quantization-, Fault- and Variability-Aware Trainingcited by 1
  30. 2025Where and How to Enhance: Discovering Bit-Width Contribution for Mixed Precision Quantizationcited by 1
  31. 2026HeRo-Q: A General Framework for Stable Low Bit Quantization via Hessian Conditioningcited by 1
  32. 2026SFMP: Fine-Grained, Hardware-Friendly and Search-Free Mixed-Precision Quantization for Large Language Modelscited by 1
  33. 2026AVQ-Attention: Adaptive Vector-Quantized Attentioncited by 1
  34. 2023Modular Quantization-Aware Training for 6D Object Pose Estimationcited by 0
  35. 2023SONIQ: System-Optimized Noise-Injected Ultra-Low-Precision Quantization with Full-Precision Paritycited by 0
  36. 2024RTF-Q: Efficient Unsupervised Domain Adaptation with Retraining-free Quantizationcited by 0
  37. 2025Precision Neural Network Quantization via Learnable Adaptive Modulescited by 0
  38. 2025CHORD: Customizing Hybrid-precision On-device Model for Sequential Recommendation with Device-cloud Collaborationcited by 0

50 most cited of 53 All 53 (JSON)