Teachable Machine

Teachable Machine is an easy-to-use web-based tool for creating machine learning models with no expertise or coding required. It can be used to recognize images, sounds, and poses, by gathering and grouping examples into classes that you want the computer to learn. Training a model is fast, and you can instantly test it out to see if it can correctly classify new examples. Models can be exported for use in sites, apps, and more, and you can choose to use it entirely on-device without any webcam or microphone data leaving your computer. Teachable Machine is compatible with TensorFlow, ML5.js, p5.js, Coral, Framer, node.js, Glitch, and Arduino, and you can use files or capture examples live. Tutorials are available to help users get started, and there are plenty of projects that have been made with Teachable Machine, such as a DIY experiment connecting Arduino and Teachable Machine, an assistive communication device, a game controller using a webcam and a piece of paper, and a physical machine that you can teach to rapidly recognize and sort objects. It can even be used to create a video game controller. It is a powerful tool for creating machine learning models with a wide range of applications.

7.2Expert Score

Awesome

Teachable Machine is an easy-to-use web-based tool for creating machine learning models with no expertise or coding required. It can be used to recognize images, sounds, and poses, by gathering and grouping examples into classes that you want the computer to learn.

Design

6.4

Easy to use

5.7

Price

6.8

Features

5.9

Accuracy

6.3

PROS

• Web-based tool
• No coding required
• Can recognize images, sounds, poses
• Fast model training
• Instant testing available
• Export models for use
• On-device usage option
• Compatible with TensorFlow
• Compatible with ML5.js
• Compatible with p5.js
• Compatible with Coral
• Compatible with Framer
• Compatible with node.js
• Compatible with Glitch
• Compatible with Arduino
• Can use files or live input
• Available tutorials
• Can create game controller
• Wide range of applications
• Can classify audio
• Can classify body positions
• Several project examples
• Models work with JavaScript
• Export to different formats
• Available lessons & activities
• Train models with own data
• Supports file and live capturing
• Several supported output formats

CONS

• No coding required limits customization
• Requires substantial data input
• Limited to three modalities: image, sound, pose
• Compatibility issues with other platforms
• Dependent on user provided class grouping
• Models may lack complexity
• Errors in classification not easily rectified
• No advanced analytics
• Limited export formats
• No user community for troubleshooting