paper-with-me

홈 › Papers

Fast Adjustable Threshold For Uniform Neural Network Quantization (Winning solution of LPIRC-II)

2018-12-19 · Alexander Goncharenko, Andrey Denisov, Sergey Alyamkin, Evgeny Terentev

Neural network quantization procedure is the necessary step for porting of neural networks to mobile devices. Quantization allows accelerating the inference, reducing memory consumption and model size. It can be performed without fine-tuning using calibration procedure (calculation of parameters necessary for quantization), or it is possible to train the network with quantization from scratch. Training with quantization from scratch on the labeled data is rather long and resource-consuming procedure. Quantization of network without fine-tuning leads to accuracy drop because of outliers which appear during the calibration. In this article we suggest to simplify the quantization procedure significantly by introducing the trained scale factors for quantization thresholds. It allows speeding up the process of quantization with fine-tuning up to 8 epochs as well as reducing the requirements to the set of train images. By our knowledge, the proposed method allowed us to get the first public available quantized version of MNAS without significant accuracy reduction - 74.8% vs 75.3% for original full-precision network. Model and code are ready for use and available at: https://github.com/agoncharenko1992/FAT-fast_adjustable_threshold.

📄 PDF Abstract BibTeX arXiv:1812.07872

Code (4)

NervanaSystems/distiller 공식 구현 pytorch
agoncharenko1992/FAT-fast-adjustable-threshold 공식 구현 tf
agoncharenko1992/FAT-fast_adjustable_threshold 공식 구현 tf
maheshkaran/Nervana-Distiller pytorch

Tasks

Quantization

Similar Papers 제목 키워드 기반

Nonuniform-to-Uniform Quantization: Towards Accurate Quantization via Generalized Straight-Through Estimation

2021-11-29 · CVPR 2022 1 · Zechun Liu, Kwang-Ting Cheng, Dong Huang, Eric Xing 외

The nonuniform quantization strategy for compressing neural networks usually achieves better performance than its counterpart, i.e., uniform strategy, due to its superior representational capacity. However, many nonunifo…

Quantization

$\text{Log}_\text{b}$Quant: Quantizing Language Models in Logarithmic Space

2026-07-01 · Jeremias Bohn, Tizian Dippold, Mahdi Koubaa, Elias R. Wahl 외 arxiv

Quantization has become an invaluable tool to reduce memory requirements and inference speed of modern language models, in particular to make them available for consumer setups and edge devices. While previous work has p…

Received Power Maximization Using Nonuniform Discrete Phase Shifts for RISs With a Limited Phase Range

2024-06-23 · Dogan Kutay Pekcan, Hongyi Liao, Ender Ayanoglu

To maximize the received power at a user equipment, the problem of optimizing a reconfigurable intelligent surface (RIS) with a limited phase range R < 2{\pi} and nonuniform discrete phase shifts with adjustable gains is…

2kQuantization

RoPE-Aware Bit Allocation for KV-Cache Quantization

2026-06-23 · Fengfeng Liang, Yuechen Zhang, Jiaya Jia arxiv

Existing low-bit KV-cache quantizers often treat each cached key as a flat vector. Under RoPE, however, a key's contribution to a future attention logit decomposes into a position-dependent sum over two-dimensional frequ…

On Uniform Scalar Quantization for Learned Image Compression

2023-09-29 · Haotian Zhang, Li Li, Dong Liu

Learned image compression possesses a unique challenge when incorporating non-differentiable quantization into the gradient-based training of the networks. Several quantization surrogates have been proposed to fulfill th…

Image CompressionQuantization