paper-with-me

홈 › Papers

AdaBits: Neural Network Quantization with Adaptive Bit-Widths

2019-12-20 · CVPR 2020 6 · Qing Jin, Linjie Yang, Zhenyu Liao

Deep neural networks with adaptive configurations have gained increasing attention due to the instant and flexible deployment of these models on platforms with different resource budgets. In this paper, we investigate a novel option to achieve this goal by enabling adaptive bit-widths of weights and activations in the model. We first examine the benefits and challenges of training quantized model with adaptive bit-widths, and then experiment with several approaches including direct adaptation, progressive training and joint training. We discover that joint training is able to produce comparable performance on the adaptive model as individual models. We further propose a new technique named Switchable Clipping Level (S-CL) to further improve quantized models at the lowest bit-width. With our proposed techniques applied on a bunch of models including MobileNet-V1/V2 and ResNet-50, we demonstrate that bit-width of weights and activations is a new option for adaptively executable deep neural networks, offering a distinct opportunity for improved accuracy-efficiency trade-off as well as instant adaptation according to the platform constraints in real-world applications.

📄 PDF Abstract BibTeX arXiv:1912.09666

Code (1)

deJQK/AdaBits pytorch

Tasks

Quantization

Similar Papers 제목 키워드 기반

Bitwidth-Adaptive Quantization-Aware Neural Network Training: A Meta-Learning Approach

2022-07-20 · Jiseok Youn, Jaehun Song, Hyung-Sin Kim, Saewoong Bahk

Deep neural network quantization with adaptive bitwidths has gained increasing attention due to the ease of model deployment on various platforms with different resource budgets. In this paper, we propose a meta-learning…

Few-Shot LearningMeta-LearningQuantization

Arbitrary Bit-width Network: A Joint Layer-Wise Quantization and Adaptive Inference Approach

2022-04-21 · Chen Tang, Haoyu Zhai, Kai Ouyang, Zhi Wang 외

Conventional model quantization methods use a fixed quantization scheme to different data samples, which ignores the inherent "recognition difficulty" differences between various samples. We propose to feed different dat…

Quantization

AdaQAT: Adaptive Bit-Width Quantization-Aware Training

2024-04-22 · Cédric Gernigon, Silviu-Ioan Filip, Olivier Sentieys, Clément Coggiola 외

Large-scale deep neural networks (DNNs) have achieved remarkable success in many application scenarios. However, high computational complexity and energy costs of modern DNNs make their deployment on edge devices challen…

Quantization

MixQuant: Adaptive Mixed-Precision Quantization for Large Language Models

2026-07-25 · Ashitabh Misra, Madhav Agrawal, Arham Jain, Tarek Abdelzaher arxiv

Mixed-precision quantization improves the accuracy of post-training quantization by allocating higher bitwidths to sensitive layers, but existing methods solve the allocation for a single fixed memory budget. In practice…

Improved Techniques for Quantizing Deep Networks with Adaptive Bit-Widths

2021-03-02 · Ximeng Sun, Rameswar Panda, Chun-Fu Chen, Naigang Wang 외

Quantizing deep networks with adaptive bit-widths is a promising technique for efficient inference across many devices and resource constraints. In contrast to static methods that repeat the quantization process and trai…

image-classificationImage ClassificationQuantizationTransfer Learning+1