The DOI matches its Crossref record.
TinyML Benchmark: Executing Fully Connected Neural Networks on Commodity Microcontrollers
Open accessAbstract
Recent advancements in the field of ultra-low-power machine learning (TinyML) promises to unlock an entirely new class of edge applications. However, continued progress is restrained by the lack of benchmarking Machine Learning (ML) models on TinyML hardware, which is fundamental to this field reaching maturity. In this paper, we designed 3 types of fully connected Neural Networks (NNs), trained each NN using 10 datasets (produces 30 NNs), and present the benchmark by reporting the onboard model performance on 7 popular MCU-boards (similar boards are used to design TinyML hardware). We open-sourced and made the complete benchmark results freely available online1to enable the TinyML community researchers and developers to systematically compare, evaluate, and improve various aspects during the design phase of ML-powered IoT hardware. 1Trained TFLite models, complete benchmark results, and more details on chosen MCU boards, datasets, NNs are available at https://github.com/bharathsudharsan/TinyML-Benchmark-NNs-on-MCUs
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The study presents an analysis of 30 neural networks on 7 different embedded boards with minimal power consumption settings for IoT.
“The study presents an analysis of 30 neural networks on 7 different embedded boards with minimal power consumption settings for IoT.”
From A review on TinyML: State-of-the-art and prospects (2022) · cites this paper as Sudharsan et al., 2021
Sudharsan et al. (2021) devised an alternative benchmark with 3 types of fully connected neural networks, available at Benchmark-NN (2021).
“In (Sudharsan et al., 2021), an alternative benchmark is devised with 3 types of fully connected neural networks, available at (Benchmark-NN, 2021).”
From A review on TinyML: State-of-the-art and prospects (2022) · cites this paper as Sudharsan et al., 2021
The alternative benchmark provides 10 datasets with various feature dimensions and class counts.
“The benchmark provides 10 datasets with various feature dimensions and class counts.”
From A review on TinyML: State-of-the-art and prospects (2022) · cites this paper as Sudharsan et al., 2021
There exist 14 popular hardware boards with various flash, SRAM, and clock speeds where such benchmarks can be tested.
“There exist 14 popular hardware boards with various flash, SRAM, and clock speeds where such benchmarks can be tested.”
From A review on TinyML: State-of-the-art and prospects (2022) · cites this paper as Sudharsan et al., 2021
Citations
Cites
No references could be resolved yet.
Cited by
Sorted by citations
A review on TinyML: State-of-the-art and prospects
Ray, P. P. (2022) · Journal of King Saud University - Computer and Information Sciences · doi:10.1016/j.jksuci.2021.11.019 · Cited by 460
Metadata
- Authors
- Sudharsan, B., Salerno, S., Nguyen, D.-D., Yahya, M., Wahid, A., Yadav, P., Breslin, J. G., & Ali, M. I.
- Affiliations
- Ollscoil na Gaillimhe – University of Galway (IE)
- Dublin City University (IE)
- ORCID
- Venue
- 2021 IEEE 7th World Forum on Internet of Things (WF-IoT)
- Publisher
- IEEE
- Published
- 2021-06-14
- Pages
- 883-884
- OpenAlex
- W3174833680
- Cited by
- 60
- References
- 0
- Citations by year
- 2025: 6 · 2024: 10 · 2023: 16 · 2022: 20 · 2021: 7
- FWCI
- 12.25
- Type
- conference paper
- Language
- en
- Open access
- green
- hdl.handle.net
- Licence
- Locations
- IEEE World Forum on Internet of Things (WF-IoT) (publishedVersion)
- ARAN (University of Galway Research Repository) (Ollscoil na Gaillimhe – University of Galway) (acceptedVersion, cc-by)
- Arrow@dit (Dublin Institute of Technology) (submittedVersion)
- University of Galway Research Repository (submittedVersion, cc-by)
- Research Repository UCD (University College Dublin) (cc-by)
- Funders
- Science Foundation Ireland
- Topics
- Physical Sciences › Computer Science › Computer Networks and Communications › IoT and Edge/Fog Computing
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Hardware and Architecture › Energy Efficiency in Computing
- Keywords
- TinyML, fully connected neural network, microcontroller, benchmarking, edge computing, IoT hardware, TensorFlow Lite, model evaluation
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
