# On-Device Training of Fully Quantized Deep Neural Networks on Cortex-M Microcontrollers

arXiv: https://arxiv.org/abs/2407.10734 · HTML: https://mlsanity.com/paper/2407.10734/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 23 of its 45 references that are in the map.
- First posted on arXiv 15 Jul 2024.

## Builds on (most cited first, 23 of 23)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms](https://mlsanity.com/paper/1708.07747/) (2017)
- [Searching for MobileNetV3](https://mlsanity.com/paper/1905.02244/) (2019)
- [Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference](https://mlsanity.com/paper/1712.05877/) (2017)
- [Training Deep Nets with Sublinear Memory Cost](https://mlsanity.com/paper/1604.06174/) (2016)
- [Speech Commands: A Dataset for Limited-Vocabulary Speech Recognition](https://mlsanity.com/paper/1804.03209/) (2018)
- [Quantizing deep convolutional networks for efficient inference: A whitepaper](https://mlsanity.com/paper/1806.08342/) (2018)
- [Deep Learning for Classical Japanese Literature](https://mlsanity.com/paper/1812.01718/) (2018)
- [MCUNet: Tiny Deep Learning on IoT Devices](https://mlsanity.com/paper/2007.10319/) (2020)
- [Training Deep Neural Networks with 8-bit Floating Point Numbers](https://mlsanity.com/paper/1812.08011/) (2018)
- [On-Device Training Under 256KB Memory](https://mlsanity.com/paper/2206.15472/) (2022)
- [Training BatchNorm and Only BatchNorm: On the Expressive Power of Random Features in CNNs](https://mlsanity.com/paper/2003.00152/) (2020)
- [Memory-Efficient Backpropagation Through Time](https://mlsanity.com/paper/1606.03401/) (2016)
- [TinyOL: TinyML with Online-Learning on Microcontrollers](https://mlsanity.com/paper/2103.08295/) (2021)
- [Visual Wake Words Dataset](https://mlsanity.com/paper/1906.05721/) (2019)
- [Tiny Machine Learning: Progress and Futures](https://mlsanity.com/paper/2403.19076/) (2024)
- [E2-Train: Training State-of-the-art CNNs with Over 80% Energy Savings](https://mlsanity.com/paper/1910.13349/) (2019)
- [K for the Price of 1: Parameter-efficient Multi-task and Transfer Learning](https://mlsanity.com/paper/1810.10703/) (2018)
- [Mandheling: Mixed-Precision On-Device DNN Training with DSP Offloading](https://mlsanity.com/paper/2206.07509/) (2022)
- [POET: Training Neural Networks on Tiny Devices with Integrated Rematerialization and Paging](https://mlsanity.com/paper/2207.07697/) (2022)
- [Dynamic Sparse Graph for Efficient Deep Learning](https://mlsanity.com/paper/1810.00859/) (2018)
- [Energy-efficient Deployment of Deep Learning Applications on Cortex-M based Microcontrollers using Deep Compression](https://mlsanity.com/paper/2205.10369/) (2022)
- [OpTC -- A Toolchain for Deployment of Neural Networks on AURIX TC3xx Microcontrollers](https://mlsanity.com/paper/2404.15833/) (2024)

## Built on by (most cited first, 5 of 5)

- [Neural Network Quantization for Microcontrollers: A Comprehensive Survey of Methods, Platforms, and Applications](https://mlsanity.com/paper/2508.15008/) (2025)
- [FORGE: Forward-Only Test-Time Adaptation for Integer-Only Vision Models on Microcontrollers](https://mlsanity.com/paper/2609.01683/) (2026)
- [What changes after deployment? A survey on On-device Learning in TinyML](https://mlsanity.com/paper/2605.31226/) (2026)
- [Hardware-Software Co-Design for Float16 On-Device Training on RISC-V Single-Core](https://mlsanity.com/paper/2607.21130/) (2026)
- [Design-Time Optimization of Deep Neural Networks for Intermittent Learning on Microcontrollers](https://mlsanity.com/paper/2608.03589/) (2026)

## Abstract (arXiv)

On-device training of DNNs allows models to adapt and fine-tune to newly collected data or changing domains while deployed on microcontroller units (MCUs). However, DNN training is a resource-intensive task, making the implementation and execution of DNN training algorithms on MCUs challenging due to low processor speeds, constrained throughput, limited floating-point support, and memory constraints. In this work, we explore on-device training of DNNs for Cortex-M MCUs. We present a method that enables efficient training of DNNs completely in place on the MCU using fully quantized training (FQT) and dynamic partial gradient updates. We demonstrate the feasibility of our approach on multiple vision and time-series datasets and provide insights into the tradeoff between training accuracy, memory overhead, energy, and latency on real hardware.