LogoLogo
HomeAPI & SDKsProjectsForumStudio
  • Getting started
    • For beginners
    • For ML practitioners
    • For embedded engineers
  • Frequently asked questions (FAQ)
  • Tutorials
    • End-to-end tutorials
      • Computer vision
        • Image classification
        • Object detection
          • Object detection with bounding boxes
          • Detect objects with centroid (FOMO)
        • Visual anomaly detection
        • Visual regression
      • Audio
        • Sound recognition
        • Keyword spotting
      • Time-series
        • Motion recognition + anomaly detection
        • Regression + anomaly detection
        • HR/HRV
        • Environmental (Sensor fusion)
    • Data
      • Data ingestion
        • Collecting image data from the Studio
        • Collecting image data with your mobile phone
        • Collecting image data with the OpenMV Cam H7 Plus
        • Using the Edge Impulse Python SDK to upload and download data
        • Trigger connected board data sampling
        • Ingest multi-labeled data using the API
      • Synthetic data
        • Generate audio datasets using Eleven Labs
        • Generate image datasets using Dall-E
        • Generate keyword spotting datasets using Google TTS
        • Generate physics simulation datasets using PyBullet
        • Generate timeseries data with MATLAB
      • Labeling
        • Label audio data using your existing models
        • Label image data using GPT-4o
      • Edge Impulse Datasets
    • Feature extraction
      • Building custom processing blocks
      • Sensor fusion using embeddings
    • Machine learning
      • Classification with multiple 2D input features
      • Visualize neural networks decisions with Grad-CAM
      • Sensor fusion using embeddings
      • FOMO self-attention
    • Inferencing & post-processing
      • Count objects using FOMO
      • Continuous audio sampling
      • Multi-impulse (C++)
      • Multi-impulse (Python)
    • Lifecycle management
      • CI/CD with GitHub Actions
      • Data aquisition from S3 object store - Golioth on AI
      • OTA model updates
        • with Arduino IDE (for ESP32)
        • with Arduino IoT Cloud
        • with Blues Wireless
        • with Docker on Allxon
        • with Docker on Balena
        • with Docker on NVIDIA Jetson
        • with Espressif IDF
        • with Nordic Thingy53 and the Edge Impulse app
        • with Particle Workbench
        • with Zephyr on Golioth
    • API examples
      • Customize the EON Tuner
      • Ingest multi-labeled data using the API
      • Python API bindings example
      • Running jobs using the API
      • Trigger connected board data sampling
    • Python SDK examples
      • Using the Edge Impulse Python SDK to run EON Tuner
      • Using the Edge Impulse Python SDK to upload and download data
      • Using the Edge Impulse Python SDK with Hugging Face
      • Using the Edge Impulse Python SDK with SageMaker Studio
      • Using the Edge Impulse Python SDK with TensorFlow and Keras
      • Using the Edge Impulse Python SDK with Weights & Biases
    • Expert network projects
  • Edge Impulse Studio
    • Organization hub
      • Users
      • Data campaigns
      • Data
        • Cloud data storage
      • Data pipelines
      • Data transformation
        • Transformation blocks
      • Upload portals
      • Custom blocks
        • Custom AI labeling blocks
        • Custom deployment blocks
        • Custom learning blocks
        • Custom processing blocks
        • Custom synthetic data blocks
        • Custom transformation blocks
      • Health reference design
        • Synchronizing clinical data with a bucket
        • Validating clinical data
        • Querying clinical data
        • Transforming clinical data
    • Project dashboard
      • Select AI hardware
    • Devices
    • Data acquisition
      • Uploader
      • Data explorer
      • Data sources
      • Synthetic data
      • Labeling queue
      • AI labeling
      • CSV Wizard (time-series)
      • Multi-label (time-series)
      • Tabular data (pre-processed & non-time-series)
      • Metadata
      • Auto-labeler | deprecated
    • Impulses
    • EON Tuner
      • Search space
    • Processing blocks
      • Audio MFCC
      • Audio MFE
      • Audio Syntiant
      • Flatten
      • HR/HRV features
      • Image
      • IMU Syntiant
      • Raw data
      • Spectral features
      • Spectrogram
      • Custom processing blocks
      • Feature explorer
    • Learning blocks
      • Anomaly detection (GMM)
      • Anomaly detection (K-means)
      • Classification
      • Classical ML
      • Object detection
        • MobileNetV2 SSD FPN
        • FOMO: Object detection for constrained devices
      • Object tracking
      • Regression
      • Transfer learning (images)
      • Transfer learning (keyword spotting)
      • Visual anomaly detection (FOMO-AD)
      • Custom learning blocks
      • Expert mode
      • NVIDIA TAO | deprecated
    • Retrain model
    • Live classification
    • Model testing
    • Performance calibration
    • Deployment
      • EON Compiler
      • Custom deployment blocks
    • Versioning
    • Bring your own model (BYOM)
    • File specifications
      • deployment-metadata.json
      • ei-metadata.json
      • ids.json
      • parameters.json
      • sample_id_details.json
      • train_input.json
  • Tools
    • API and SDK references
    • Edge Impulse CLI
      • Installation
      • Serial daemon
      • Uploader
      • Data forwarder
      • Impulse runner
      • Blocks
      • Himax flash tool
    • Edge Impulse for Linux
      • Linux Node.js SDK
      • Linux Go SDK
      • Linux C++ SDK
      • Linux Python SDK
      • Flex delegates
      • Rust Library
    • Rust Library
    • Edge Impulse Python SDK
  • Run inference
    • C++ library
      • As a generic C++ library
      • On Android
      • On your desktop computer
      • On your Alif Ensemble Series Device
      • On your Espressif ESP-EYE (ESP32) development board
      • On your Himax WE-I Plus
      • On your Raspberry Pi Pico (RP2040) development board
      • On your SiLabs Thunderboard Sense 2
      • On your Spresense by Sony development board
      • On your Syntiant TinyML Board
      • On your TI LaunchPad using GCC and the SimpleLink SDK
      • On your Zephyr-based Nordic Semiconductor development board
    • Arm Keil MDK CMSIS-PACK
    • Arduino library
      • Arduino IDE 1.18
    • Cube.MX CMSIS-PACK
    • Docker container
    • DRP-AI library
      • DRP-AI on your Renesas development board
      • DRP-AI TVM i8 on Renesas RZ/V2H
    • IAR library
    • Linux EIM executable
    • OpenMV
    • Particle library
    • Qualcomm IM SDK GStreamer
    • WebAssembly
      • Through WebAssembly (Node.js)
      • Through WebAssembly (browser)
    • Edge Impulse firmwares
    • Hardware specific tutorials
      • Image classification - Sony Spresense
      • Audio event detection with Particle boards
      • Motion recognition - Particle - Photon 2 & Boron
      • Motion recognition - RASynBoard
      • Motion recognition - Syntiant
      • Object detection - SiLabs xG24 Dev Kit
      • Sound recognition - TI LaunchXL
      • Keyword spotting - TI LaunchXL
      • Keyword spotting - Syntiant - RC Commands
      • Running NVIDIA TAO models on the Renesas RA8D1
      • Two cameras, two models - running multiple object detection models on the RZ/V2L
  • Edge AI Hardware
    • Overview
    • Production-ready
      • Advantech ICAM-540
      • Seeed SenseCAP A1101
      • Industry reference design - BrickML
    • MCU
      • Ambiq Apollo4 family of SoCs
      • Ambiq Apollo510
      • Arducam Pico4ML TinyML Dev Kit
      • Arduino Nano 33 BLE Sense
      • Arduino Nicla Sense ME
      • Arduino Nicla Vision
      • Arduino Portenta H7
      • Blues Wireless Swan
      • Espressif ESP-EYE
      • Himax WE-I Plus
      • Infineon CY8CKIT-062-BLE Pioneer Kit
      • Infineon CY8CKIT-062S2 Pioneer Kit
      • Nordic Semi nRF52840 DK
      • Nordic Semi nRF5340 DK
      • Nordic Semi nRF9160 DK
      • Nordic Semi nRF9161 DK
      • Nordic Semi nRF9151 DK
      • Nordic Semi nRF7002 DK
      • Nordic Semi Thingy:53
      • Nordic Semi Thingy:91
      • Open MV Cam H7 Plus
      • Particle Photon 2
      • Particle Boron
      • RAKwireless WisBlock
      • Raspberry Pi RP2040
      • Renesas CK-RA6M5 Cloud Kit
      • Renesas EK-RA8D1
      • Seeed Wio Terminal
      • Seeed XIAO nRF52840 Sense
      • Seeed XIAO ESP32 S3 Sense
      • SiLabs Thunderboard Sense 2
      • Sony's Spresense
      • ST B-L475E-IOT01A
      • TI CC1352P Launchpad
    • MCU + AI accelerators
      • Alif Ensemble
      • Arduino Nicla Voice
      • Avnet RASynBoard
      • Seeed Grove - Vision AI Module
      • Seeed Grove Vision AI Module V2 (WiseEye2)
      • Himax WiseEye2 Module and ISM Devboard
      • SiLabs xG24 Dev Kit
      • STMicroelectronics STM32N6570-DK
      • Synaptics Katana EVK
      • Syntiant Tiny ML Board
    • CPU
      • macOS
      • Linux x86_64
      • Raspberry Pi 4
      • Raspberry Pi 5
      • Texas Instruments SK-AM62
      • Microchip SAMA7G54
      • Renesas RZ/G2L
    • CPU + AI accelerators
      • AVNET RZBoard V2L
      • BrainChip AKD1000
      • i.MX 8M Plus EVK
      • Digi ConnectCore 93 Development Kit
      • MemryX MX3
      • MistyWest MistySOM RZ/V2L
      • Qualcomm Dragonwing RB3 Gen 2 Dev Kit
      • Renesas RZ/V2L
      • Renesas RZ/V2H
      • IMDT RZ/V2H
      • Texas Instruments SK-TDA4VM
      • Texas Instruments SK-AM62A-LP
      • Texas Instruments SK-AM68A
      • Thundercomm Rubik Pi 3
    • GPU
      • Advantech ICAM-540
      • NVIDIA Jetson
      • Seeed reComputer Jetson
    • Mobile phone
    • Porting guide
  • Integrations
    • Arduino Machine Learning Tools
    • AWS IoT Greengrass
    • Embedded IDEs - Open-CMSIS
    • NVIDIA Omniverse
    • Scailable
    • Weights & Biases
  • Tips & Tricks
    • Combining impulses
    • Increasing model performance
    • Optimizing compute time
    • Inference performance metrics
  • Concepts
    • Glossary
    • Course: Edge AI Fundamentals
      • Introduction to edge AI
      • What is edge computing?
      • What is machine learning (ML)?
      • What is edge AI?
      • How to choose an edge AI device
      • Edge AI lifecycle
      • What is edge MLOps?
      • What is Edge Impulse?
      • Case study: Izoelektro smart grid monitoring
      • Test and certification
    • Data engineering
      • Audio feature extraction
      • Motion feature extraction
    • Machine learning
      • Data augmentation
      • Evaluation metrics
      • Neural networks
        • Layers
        • Activation functions
        • Loss functions
        • Optimizers
          • Learned optimizer (VeLO)
        • Epochs
    • What is embedded ML, anyway?
    • What is edge machine learning (edge ML)?
Powered by GitBook
On this page
  • Getting started with EIM artifacts
  • EIM artifact overview
  • Troubleshooting
  • Advanced Usage

Was this helpful?

Export as PDF
  1. Run inference

Linux EIM executable

PreviousIAR libraryNextOpenMV

Last updated 6 months ago

Was this helpful?

"Edge Impulse Model" (EIM) files are native Linux and macOS binary applications that contains your full impulse created in Edge Impulse Studio. The impulse consists of the signal processing block(s) along with any learning and anomaly block(s) you added and trained. EIM files are compiled for your particular system architecture and are used to run inference natively on your system.

Source code for EIM can be found .

Getting started with EIM artifacts

To download and use an EIM artifact, you will need to first install the Edge Impulse CLI tool suite by .

From there, use the edge-impulse-linux-runner tool to download the .eim file (named modelfile.eim):

edge-impulse-linux-runner --download modelfile.eim

Note that the first time you call this tool, it will ask you to log into your Edge Impulse account and select a project. Subsequent calls will use your cached credentials and assume your selected project. To select a different project, you will need to log in again by using the --clean argument:

edge-impulse-linux-runner --clean --download modelfile.eim

Note that you can also download a .eim file for your system from Edge Impulse Studio. The various .eim deployment options are listed in the Deployment page. Search for "Linux" or "macOS" to see the executables listed as possible targets for deployment.

From there, you can use one of the supported high-level language SDKs to perform inference locally. These SDKs run the .eim executable as its own process in the background and provide an interface for you to use in your application. See the following guides to get started using these SDKs:

The interface, installed as a library for your language, allows you to send raw data to the EIM process and receive inference results.

EIM artifact overview

The EIM file runs as a native Linux application. JSON data is passed to and from the EIM program using Unix-like sockets or standard input/output (stdio). The high-level architecture and data flow is shown below.

In general, you use the edge-impulse-linux-runner application to download and interact with the EIM model. When you request to download the EIM file, you will be asked to log into your Edge Impulse Studio account. Studio will compile the .eim binary from a given project for your particular computer architecture.

From there, you can use the edge-impulse-linux-runner to interact with the EIM file. This includes sending raw samples and receiving inference results. Edge Impulse maintains high-level programming language interfaces (e.g. Python, Node.JS, Go) for you to use to interact with the runner process. From this, you can develop applications to natively perform inference using your impulse trained on Edge Impulse Studio.

The high-level language SDKs talk to the model through the IPC layer to run inference, so these SDKs are very thin and just need the ability to spawn a binary.

Troubleshooting

Wrong OS bits

If you encounter the following error message when trying to run a .eim file, it likely means the .eim file was compiled for the incorrect system architecture:

Failed to run impulse Error: Unsupported architecture "<some-architecture>"

Often, it means that you are attempting to deploy the .eim file to a 32-bit operating system running on a 64-bit CPU. You can check your CPU's bit width with the following commands:

uname -m
uname -a

These should show you if you are running on a 32-bit CPU (e.g. x86) or a 64-bit CPU (e.g. x86_64 or aarch64).

Next, check your operating system bit width:

getconf LONG_BIT

This will return 32 for 32 bits or 64 for 64 bits. When you use the runner tool to download an .eim file, it will provide the Edge Impulse servers information about your hardware architecture. If you are using a 64-bit CPU, you will receive an .eim file compiled for 64-bit systems. Such an executable will not run on 32-bit operating systems.

To correct this, make sure that you are running a 32-bit OS on 32-bit hardware or a 64-bit OS on 64-bit hardware only.

Advanced Usage

If you would like to modify, debug, or interface with EIM artifacts on a lower level, these notes should help you get started.

JSON protocol

JSON frames are used to pass data to and from the EIM executable running on your system. They can be transferred via Unix-like sockets (what edge-impulse-linux-runner uses) or standard input/output (stdio). Example frames are given below.

An example dataflow of how such JSON messages are passed between the runner and EIM executable is shown below:

If the EIM executable encounters an error, you might see something like this:

Runner to EIM JSON examples

Basic "hello" frame:

{
	"hello": 1, // always 1 (version number)
	"id": 1 // ID should be incremented with each message
}

Sending raw features to be classified:

{
	"classify": [1, 2, ..., 3],
	"id": 2,
	"debug": false // OPTIONAL, enables debug output from EIM
}

EIM to runner JSON examples

Model info:

{
   "id": 1,
   "model_parameters": {
      "axis_count": 3,
      "frequency": 62.5,
      "has_anomaly": 1,
      "image_channel_count": 0,
      "image_input_frames": 0,
      "image_input_height": 0,
      "image_input_width": 0,
      "input_features_count": 375,
      "interval_ms": 16,
      "label_count": 4,
      "labels": [
         "idle",
         "snake",
         "updown",
         "wave"
      ],
      "model_type": "classification",
      "sensor": 2,
      "slice_size": 31,
      "use_continuous_mode": false
   },
   "project": {
      "deploy_version": 411,
      "id": 1,
      "name": "Continuous gestures",
      "owner": "EdgeImpulse Inc."
   },
   "success": true
}

Classification inference result:

{
	"id": 1, // ID of the 'classify' message
	"success": true,
	"result": {
		"classification": {
			"up": 0.98,
			"down": 0.00,
			"wave": 0.02
		}
	},
	"timing": {
		"dsp": 0, // miliseconds of the DSP processing
		"classification": 0, // miliseconds of the NN processing
		"anomaly": 0, // miliseconds of the anomaly block processing
		"json": 0, // miliseconds spent on the JSON parsing
		"stdin": 0 // miliseconds sepnt on the STDIN processing
	}
}

Object detection inference result:

{
	"id": 1, // ID of the 'classify' message
	"success": true,
	"result": {
		"bounding_boxes": [
		    {
		      "height": 8,
	        "label": "beer",
	        "value": 0.5000159740447998,
	        "width": 8,
	        "x": 8,
	        "y": 40
	     },
	     {
		     "height": 8,
	       "label": "beer",
	       "value": 0.5000159740447998,
	       "width": 8,
	       "x": 72,
	       "y": 40
		  }
		]
	},
	"timing": {
		"dsp": 0, // miliseconds of the DSP processing
		"classification": 0, // miliseconds of the NN processing
		"anomaly": 0, // miliseconds of the anomaly block processing
		"json": 0, // miliseconds spent on the JSON parsing
		"stdin": 0 // miliseconds sepnt on the STDIN processing
	}
}

Error frame:

{
	"success": false,
	"error": "Error message",
	"id": 1 // OPTIONAL, ID of the message that generated error
}

Debugging with stdout

You can run the .eim file as any other binary executable. Instead of using sockets to send and receive JSON frames manually, you can connect the runner to the EIM's socket to use your shell's standard input/output. To do that, open a terminal and run your .eim file as an executable, giving it a temporary socket as an argument:

./modelfile.eim /tmp/hello.sock

Open another terminal and start the impulse runner, connecting it to your temporary socket:

edge-impulse-linux-runner --model-file /tmp/hello.sock

In addition to using stdout, you can also use common debugger tools with the runner to assist in debugging the EIM program. To use LLDB, run:

lldb -o run ./modelfile.eim /tmp/hello.sock

For GDB:

gdb --args ./modelfile.eim /tmp/hello.sock

If you would like to compile the .eim file manually, please refer to .

EIM executables contain your signal processing and ML code, compiled with optimizations for your processor or GPU (e.g. NEON instructions on ARM cores) plus a very simple (over a Unix socket). By doing this, your model file is now completely self-contained, and it does not depend on anything (except glibc). As a result, you do not need specific TensorFlow versions, can avoid Python dependency hell, and will never have to worry about why you're not running at full native speed.

Linux Node.js SDK
Linux Go SDK
Linux Python SDK
IPC layer
here
following these instructions
these instructions
Download EIM file from Edge Impulse Studio
EIM artifact overview
EIM JSON data flow
EIM JSON data flow with error condition