Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Edge Impulse: An MLOps Platform for Tiny Machine Learning

Shawn Hymel, Colby Banbury, Daniel Situnayake, Alex Elium, Carl Ward, Mat Kelcey, Mathijs Baaijens, Mateusz Majchrzycki, Jenny Plunkett, David Tischler, Alessandro Grande, Louis Moreau, Dmitry Maslov, Artie Beavis, Jan Jongboom, Vijay Janapa Reddi

TL;DR

Edge Impulse addresses fragmentation and heterogeneity in TinyML by delivering an end-to-end MLOps platform with DSP-aware AutoML, extensible inference, and scalable cloud infrastructure. The approach integrates data collection, DSP preprocessing, model design, optimization (EON Compiler, EON Tuner), and deployment across diverse edge hardware to enable cross-stack co-optimization. The paper demonstrates performance benefits across devices, showcases education and open-source ecosystem benefits, and presents industry deployments (Oura Ring, SlateSafety) to illustrate real-world impact. Overall, Edge Impulse lowers the expertise and compute barriers for TinyML development, enabling portable, scalable edge ML solutions.

Abstract

Edge Impulse is a cloud-based machine learning operations (MLOps) platform for developing embedded and edge ML (TinyML) systems that can be deployed to a wide range of hardware targets. Current TinyML workflows are plagued by fragmented software stacks and heterogeneous deployment hardware, making ML model optimizations difficult and unportable. We present Edge Impulse, a practical MLOps platform for developing TinyML systems at scale. Edge Impulse addresses these challenges and streamlines the TinyML design cycle by supporting various software and hardware optimizations to create an extensible and portable software stack for a multitude of embedded systems. As of Oct. 2022, Edge Impulse hosts 118,185 projects from 50,953 developers.

Edge Impulse: An MLOps Platform for Tiny Machine Learning

TL;DR

Edge Impulse addresses fragmentation and heterogeneity in TinyML by delivering an end-to-end MLOps platform with DSP-aware AutoML, extensible inference, and scalable cloud infrastructure. The approach integrates data collection, DSP preprocessing, model design, optimization (EON Compiler, EON Tuner), and deployment across diverse edge hardware to enable cross-stack co-optimization. The paper demonstrates performance benefits across devices, showcases education and open-source ecosystem benefits, and presents industry deployments (Oura Ring, SlateSafety) to illustrate real-world impact. Overall, Edge Impulse lowers the expertise and compute barriers for TinyML development, enabling portable, scalable edge ML solutions.

Abstract

Edge Impulse is a cloud-based machine learning operations (MLOps) platform for developing embedded and edge ML (TinyML) systems that can be deployed to a wide range of hardware targets. Current TinyML workflows are plagued by fragmented software stacks and heterogeneous deployment hardware, making ML model optimizations difficult and unportable. We present Edge Impulse, a practical MLOps platform for developing TinyML systems at scale. Edge Impulse addresses these challenges and streamlines the TinyML design cycle by supporting various software and hardware optimizations to create an extensible and portable software stack for a multitude of embedded systems. As of Oct. 2022, Edge Impulse hosts 118,185 projects from 50,953 developers.
Paper Structure (38 sections, 3 figures, 5 tables)

This paper contains 38 sections, 3 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Embedded Ecosystem Challenges
  3. Device Resource Constraints
  4. Hardware Heterogeneity
  5. Software Fragmentation
  6. Co-Optimization and Cross-Stack Collaboration
  7. Overview and Design Objectives
  8. Implementation
  9. Data Collection and Analysis
  10. DSP Pipeline
  11. ML Design and Training
  12. Estimation and Evaluation
  13. Compression and Optimization
  14. Conversion and Compilation
  15. AutoML
  16. ...and 23 more sections

Figures (3)

  • Figure 1: The challenges associated with the ML Workflow and features of Edge Impulse that solve those challenges.
  • Figure 2: Screenshot showing the user's view inside an Edge Impulse project where the blocks are connected depicting the dataflow.
  • Figure 3: Screenshot of the EON Tuner. Features are annotated with color coded dotted boxes that correspond to the challenges in Figure \ref{['fig:ml-workflow']}. Purple (top right): The tuner allows users to select the target hardware, which will then inform the constraints set on the search. Blue (top left): The tuner computes the configuration's accuracy and predicts the resource consumption of the DSP and NN components. Pink (bottom): The tuner searches for optimal DSP and NN combinations and displays their configuration.