A List of Edge Computing Saas Platforms
In machine learning, you had to collect a lot of data, preprocess and build your model with coding. In order to do that, you need to learn coding in Python or C/C++ and Python libraries such as Tensorflow, Keras, PyTorch, etc. Let me tell you that, there are now a few successful SaaS platforms that you can build a machine learning model without minimum coding. You will find the list of these Saas companies and their features.
List of TinyML Companies
This is the list of TinyML SaaS websites:
Edge Impulse https://www.edgeimpulse.com/
AlwaysAI https://alwaysai.co/
Neuton AI https://neuton.ai/
Qeexo https://qeexo.com/
COMPARISON
In this part of the article, a fast comparison in terms of pricing, ease of use, supported development platforms, live testing is provided.
Pricing
Edge Impulse is available to developers for free with 20 minute training time and 4 GB or 4 hours of data limitation. For commercial license, price is not listed.
AlwaysAI is also free to developers with limitations. For commercial license, price is not listed.
NeutonAI is also free to developers with limitations. This platform utilizes Google Cloud Platform with a minimum of 100 hour training hour.
Qeexo is also free to developers with limitations: 2GB Storage, no commercial license. For commercial license, price is not listed.
Ease of Use
Edge Impulse have you install a Data forwarder on command line interface. After you collect data, the rest is easy to manage as the interface is based on Web page.
AlwaysAI needs you to install their software which works on a command line interface. AlwaysAI has ready to use models provided for supported platforms.
Qeexo also asks for Qeexo AutoML software to be installed. It also incorporates features like Edge Impulse.
Neuton AI has an interface with ready data uploading. For developers that prefer simplicity, Neuton AI is the way to go.
Supported Development Platforms
Edge Impulse:
Officially Supported MCUs:
- Arduino Nano 33 BLE Sense
- Arduino Nicla Sense ME
- Arduino Nicla Vision
- Arduino Portenta H7 + Vision Shield
- Espressif ESP32 (ESP-EYE)
- Himax WE-I Plus
- Nordic Semi nRF52840 DK
- Nordic Semi nRF5340 DK
- Nordic Semi nRF9160 DK
- Nordic Semi Thingy:53
- Nordic Semi Thingy:91
- Open MV Cam H7 Plus
- Silicon Labs xG24 Dev Kit
- Silicon Labs Thunderboard Sense 2
- Sony’s Spresense
- ST B-L475E-IOT01A
- Syntiant Tiny ML Board
- TI CC1352P Launchpad
- Raspberry Pi RP2040
- Alif Ensemble E7
Officially Supported CPU-GPU Targets
- Intel Based Macs
- Linux x86/x64
- NVIDIA Jetson Nano
- Raspberry Pi 4
Community Boards
- Seed Wio Terminal
- ArduCAM Pico4ML TinyML Dev Kit
- Blues Wireless Swan
- Ease of Use
AlwaysAI:
- Raspberry Pi 4
- Raspberry Pi 3B+
- NVIDIA Jetson Nano
- NVIDIA Jetson Xavier NX
- Inforce 6640
- Dragonboard 410c
- Intel NCS2
- CUDA
- Hailo-8
Qeexo:
- Arduino Nicla Sense ME
- Arduino Nano 33 BLE Sense
- Arduino Nano 33 IoT ST
- SensorTile.box
- STWIN Wireless Industrial Node
Neuton AI:
- Any Arduino platform
Firmware
All platforms can provide a static firmware output. If you have the level of experience, you can modify these firmware to suit your needs. Otherwise, the platforms provide firmware engineering services as well.