mlsanityData to 13 Oct 2026

MOHAQ: Multi-Objective Hardware-Aware Quantization of Recurrent Neural Networks

arXiv 2108.01192 ↗First posted 2 Aug 2021cs.LG cs.ARcited by 0

Not on the map. Why?

Abstract (arXiv)

The compression of deep learning models is of fundamental importance in deploying such models to edge devices. The selection of compression parameters can be automated to meet changes in the hardware platform and application using optimization algorithms. This article introduces a Multi-Objective Hardware-Aware Quantization (MOHAQ) method, which considers hardware efficiency and inference error as objectives for mixed-precision quantization. The proposed method feasibly evaluates candidate solutions in a large search space by relying on two steps. First, post-training quantization is applied for fast solution evaluation (inference-only search). Second, we propose the "beacon-based search" to retrain selected solutions only and use them as beacons to know the effect of retraining on other solutions. We use a speech recognition model based on Simple Recurrent Unit (SRU) using the TIMIT dataset and apply our method to run on SiLago and Bitfusion platforms. We provide experimental evaluations showing that SRU can be compressed up to 8x by post-training quantization without any significant error increase. On SiLago, we found solutions that achieve 97% and 86% of the maximum possible speedup and energy saving, with a minor increase in error. On Bitfusion, beacon-based search reduced the error gain of inference-only search by up to 4.9 percentage points.

Key facts. No citing papers in the map yet (data through 13 Oct 2026). Builds on 23 of its 55 references that are in the map. First posted on arXiv 2 Aug 2021.

Builds on

23 of 55 references.

References · 23

  1. 2020Up or Down? Adaptive Rounding for Post-Training Quantizationcited by 617
  2. 2015BinaryConnect: Training Deep Neural Networks with binary weights during propagationscited by 589
  3. 2018HAQ: Hardware-Aware Automated Quantization with Mixed Precisioncited by 439
  4. 2020ZeroQ: A Novel Zero Shot Quantization Frameworkcited by 269
  5. 2020pymoo: Multi-objective Optimization in Pythoncited by 197
  6. 2019Improving Neural Network Quantization without Retraining using Outlier Channel Splittingcited by 151
  7. 2018NetAdapt: Platform-Aware Neural Network Adaptation for Mobile Applicationscited by 146
  8. 2016Designing Energy-Efficient Convolutional Neural Networks using Energy-Aware Pruningcited by 133
  9. 2017Bit Fusion: Bit-Level Dynamically Composable Architecture for Accelerating Deep Neural Networkscited by 109
  10. 2019Loss Aware Post-training Quantizationcited by 95
  11. 2020Post-Training Piecewise Linear Quantization for Deep Neural Networkscited by 78
  12. 2017NeST: A Neural Network Synthesis Tool Based on a Grow-and-Prune Paradigmcited by 60
11 more
  1. 2018Light Gated Recurrent Units for Speech Recognitioncited by 48
  2. 2018The PyTorch-Kaldi Speech Recognition Toolkitcited by 39
  3. 2015Resiliency of Deep Neural Networks under Quantizationcited by 34
  4. 2019Optimizing Speech Recognition For The Edgecited by 32
  5. 2017DeepIoT: Compressing Deep Neural Network Structures for Sensing Systems with a Compressor-Critic Frameworkcited by 22
  6. 2018C-LSTM: Enabling Efficient LSTM using Structured Compression Techniques on FPGAscited by 21
  7. 2018FINN-L: Library Extensions and Design Trade-off Analysis for Variable Precision LSTM Networks on FPGAscited by 10
  8. 2019Recurrent Neural Networks: An Embedded Computing Perspectivecited by 9
  9. 2018E-RNN: Design Optimization for Efficient Recurrent Neural Networks in FPGAscited by 8
  10. 2018Approximate FPGA-based LSTMs under Computation Time Constraintscited by 2
  11. 2019Neural Machine Translation with 4-Bit Precision and Beyondcited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2022Sub-8-bit quantization for on-device speech recognition: a regularization-free approach
  2. 2022Sub-8-Bit Quantization Aware Training for 8-Bit Neural Network Accelerator with On-Device Speech Recognition
  3. 2023RepQ: Generalizing Quantization-Aware Training for Re-Parametrized Architectures
  4. 2025Dual Precision Quantization for Efficient and Accurate Deep Neural Networks Inference
  5. 2022Accelerating Inference and Language Model Fusion of Recurrent Neural Network Transducers via End-to-End 4-bit Quantization
  6. 2024Optimizing Large Language Models through Quantization: A Comparative Analysis of PTQ and QAT Techniques
  7. 2023A Model for Every User and Budget: Label-Free and Personalized Mixed-Precision Quantization
  8. 2026Quant.npu: Enabling Efficient Mobile NPU Inference for on-device LLMs via Fully Static Quantization