arXiv Machine Learning

Deep Microcompression: Structured Pruning and Bit-packed Quantization for Microcontrollers

Deep Microcompression (DMC) is a hardware‑aware pipeline that combines structured pruning, quantization‑aware training, and fixed‑length bit‑packing to enable deep learning inference on bare‑metal microcontrollers. The method achieves a 55.8× weight compression on LeNet‑5 while maintaining 98.77% accuracy, and produces a dependency‑free C library with deterministic latency. On the RP2040 Cortex‑M0+ microcontroller, DMC cuts binary size threefold compared to TensorFlow Lite while matching its accuracy, and it is the first documented deployment of a standard CNN on the 2 KB SRAM ATmega328P.

arXiv Machine Learning
Aug 28

Pruning Binarized Neural Networks: A Dedicated Framework and Globally Weighted Algorithms

The paper presents a PyTorch-based framework for designing and optimizing binarized neural networks, incorporating freezing and pruning mechanisms. It introduces a novel pruning method that uses a global weighting scheme to assess parameter importance across abstraction levels, achieving a 70% pruning rate on VGG11 without sacrificing accuracy—outperforming existing binarized pruning results of 41%. The framework facilitates rapid, reproducible evaluation and prototyping of state‑of‑the‑art binarized network techniques.

By Roan Rubiales, Jean Pierre David
arXiv AI
1d ago

REQAP: Resilient Weight Packing and Quantization for Edge DNN Acceleration

The paper introduces REQAP, a reliability‑aware quantized weight packing technique for systolic‑array DNN accelerators. It uses a sensitivity‑driven mixed‑precision quantization to assign layer‑wise bit‑widths, a deterministic register‑level packing strategy for SIMD‑within‑a‑register execution, and selective bit‑level protection that replicates critical MSBs into unused register space. Experiments on AlexNet, VGG‑11, and ResNet‑18 show up to 62% memory reduction, 56% fewer MAC operations, and improved accuracy resilience under fault injection compared to baseline and fully protected models.

By Mahdi Taheri, Samira Nazari, Mubassher Ansari, Ali Azarpeyvand, Mohsen Afsharchi, Maksim Jenihhin, Christian Herglotz
arXiv Machine Learning
Sep 10

Squeeze10-LLM: Squeezing LLMs' Weights by 10 Times via a Staged Mixed-Precision Quantization Method

Squeeze10-LLM is a staged mixed‑precision post‑training quantization framework that reduces 16‑bit LLM weights to an average of 1.6 bits per weight by assigning 80% of weights to 1 bit and 20% to 4 bits. It introduces Post‑Binarization Activation Robustness (PBAR), a weight significance metric that considers activation impact, and Full Information Activation Supervision (FIAS), a strategy that preserves activation information to limit error propagation. Experiments on LLaMA and LLaMA2 demonstrate that Squeeze10‑LLM achieves state‑of‑the‑art performance for sub‑2‑bit weight‑only quantization, raising average accuracy from 43% to 56% on six zero‑shot classification tasks.

By Qingcheng Zhu, Yangyang Ren, Linlin Yang, Yanjing Li, Sheng Xu, Haodong Zhu, Juan Zhang, Runqi Wang, Baochang Zhang
arXiv Computation and Language
Sep 1

Budget-Aware Compression Pipeline for Single-GPU LLM Inference: Methods, Trade-offs, and Coupling Effects

The paper presents a budget‑aware compression pipeline for deploying 70B‑parameter language models on a single NVIDIA GPU. It examines how pruning, quantization, and KV‑cache compression interact, showing that layer‑wise pruning improves weight quantization robustness and that KV‑cache sparsification complements INT8 KV quantization without harming decoding speed. Using these insights, the authors compressed a 70B model to ~33 GB, achieving ~57 tokens/s on 10k‑token prompts on an A40 while maintaining accuracy within 5% on standard benchmarks.

By Hongyu Yu, Yifei Shen