paper-with-me

홈 › Papers

Leveraging Automated Mixed-Low-Precision Quantization for tiny edge microcontrollers

2020-08-12 · Manuele Rusci, Marco Fariselli, Alessandro Capotondi, Luca Benini

The severe on-chip memory limitations are currently preventing the deployment of the most accurate Deep Neural Network (DNN) models on tiny MicroController Units (MCUs), even if leveraging an effective 8-bit quantization scheme. To tackle this issue, in this paper we present an automated mixed-precision quantization flow based on the HAQ framework but tailored for the memory and computational characteristics of MCU devices. Specifically, a Reinforcement Learning agent searches for the best uniform quantization levels, among 2, 4, 8 bits, of individual weight and activation tensors, under the tight constraints on RAM and FLASH embedded memory sizes. We conduct an experimental analysis on MobileNetV1, MobileNetV2 and MNasNet models for Imagenet classification. Concerning the quantization policy search, the RL agent selects quantization policies that maximize the memory utilization. Given an MCU-class memory bound of 2MB for weight-only quantization, the compressed models produced by the mixed-precision engine result as accurate as the state-of-the-art solutions quantized with a non-uniform function, which is not tailored for CPUs featuring integer-only arithmetic. This denotes the viability of uniform quantization, required for MCU deployments, for deep weights compression. When also limiting the activation memory budget to 512kB, the best MobileNetV1 model scores up to 68.4% on Imagenet thanks to the found quantization policy, resulting to be 4% more accurate than the other 8-bit networks fitting the same memory constraints.

📄 PDF Abstract BibTeX arXiv:2008.05124

Code (0)

등록된 구현이 없습니다.

Tasks

Quantization

Methods 이 논문이 사용한 방법론

Pointwise Convolution Pointwise Convolution is a type of convolution that uses a 1x1 kernel: a kernel that iterates through every single point. This…
ReLU How Do I Communicate to Expedia? How Do I Communicate to Expedia? – Call ☎️ +1-(888) 829 (0881) or +1-805-330-4056 or +1-805-330-4056 for Live Support & Special Travel…
Depthwise Convolution Depthwise Convolution is a type of convolution where we apply a single convolutional filter for each input channel. In the regular 2D…
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
MobileNetV1 MobileNet is a type of convolutional neural network designed for mobile and embedded vision applications. They are based on a streamlined architecture that uses depthwise…
Average Pooling 설명 없음
Convolution A convolution is a type of matrix operation, consisting of a kernel, a small matrix of weights, that slides over input data performing element-wise multiplication with the…
Sigmoid Activation 설명 없음

Similar Papers 제목 키워드 기반

FairQuant: Fairness-Aware Mixed-Precision Quantization for Medical Image Classification

2026-02-26 · Thomas Woergaard, Raghavendra Selvan arxiv

Compressing neural networks by quantizing model parameters offers useful trade-off between performance and efficiency. Methods like quantization-aware training and post-training quantization strive to maintain the downst…

Medical Image Classification

TinyMusician: On-Device Music Generation with Knowledge Distillation and Mixed Precision Quantization

2025-08-31 · Hainan Wang, Mehdi Hosseinzadeh, Reza Rawassizadeh arxiv

The success of the generative model has gained unprecedented attention in the music generation area. Transformer-based architectures have set new benchmarks for model performance. However, their practical adoption is hin…

Knowledge DistillationMusic Generation

Channel-wise Mixed-precision Assignment for DNN Inference on Constrained Edge Nodes

2022-06-17 · Matteo Risso, Alessio Burrello, Luca Benini, Enrico Macii 외

Quantization is widely employed in both cloud and edge systems to reduce the memory occupation, latency, and energy consumption of deep neural networks. In particular, mixed-precision quantization, i.e., the use of diffe…

Neural Architecture SearchQuantization

Entropy-Driven Mixed-Precision Quantization for Deep Network Design

2022-11-28 · Conference on Neural Information Processing Systems 2022 11 · Zhenhong Sun, Ce Ge, Junyan Wang, Ming Lin 외

Deploying deep convolutional neural networks on Internet-of-Things (IoT) devices is challenging due to the limited computational resources, such as limited SRAM memory and Flash storage. Previous works re-design a small …

CPUFace DetectionGPUHardware Aware Neural Architecture Search+3

BMPQ: Bit-Gradient Sensitivity Driven Mixed-Precision Quantization of DNNs from Scratch

2021-12-24 · Souvik Kundu, Shikai Wang, Qirui Sun, Peter A. Beerel 외

Large DNNs with mixed-precision quantization can achieve ultra-high compression while retaining high classification performance. However, because of the challenges in finding an accurate metric that can guide the optimiz…

QuantizationSensitivity