https://andreasref.github.io/cita/teachablemachine2.gif

Link: TeachableMachine.withgoogle.com

1) Training TM step-by-step

  1. Go to TeachableMachine.withgoogle.com

  2. Press the "get started" button

  3. Click on "image project" and select standard

  4. Add as many classes as you like and give them new names

    https://s3-us-west-2.amazonaws.com/secure.notion-static.com/a39b7505-8802-48cc-877e-faf5b64719ea/0_rCj4e12llDVwG1aH.png

  5. Start recording examples for each class

    https://s3-us-west-2.amazonaws.com/secure.notion-static.com/b755e696-d42c-43a5-a9c2-c6b4fd176467/0_ymnHiaKa5dIFkCbj.gif

    https://s3-us-west-2.amazonaws.com/secure.notion-static.com/e0505dba-0ae4-47b4-9c0c-cb5eb24b01e2/0_v5DL_GStG1cLLoRb.gif

  6. Click "train model"

    https://s3-us-west-2.amazonaws.com/secure.notion-static.com/94ce3940-78a5-4829-9f68-e244e8ada2c2/0_YtOobQkpflCEAWsE.gif

  7. Once the model is trained you can preview if the model works!

    https://s3-us-west-2.amazonaws.com/secure.notion-static.com/d3cdc278-eff2-426a-b91c-def3a97cfe52/0_4hg7fS-BPMDn-kbc.gif

<aside> 👉 Exercise 1 (15 minutes): Come up with a fun idea and train the Image Model of Teachable Machine using your webcam. Make a small video/gif documenting your model and post it on Slack in the playful-ml_2021 channel. If you finish early check out this article for tips and tricks on how to train a more robust model with Teachable Machine. Keep your Teachable Machine browser tab open (but turn off the webcam preview), we will use your model in the next exercise.

</aside>

<aside> ❓ Follow up discussion: Was it easy? Was it hard? What worked? What did not work? Walk-trough of article on tips and tricks for training a model on the fly

</aside>

2) Teachable Machine + P5JS

Setup an account with the p5js online editor (optional) p5js webpage - p5js examples - p5js reference

https://s3-us-west-2.amazonaws.com/secure.notion-static.com/7fe8d18b-8625-4e53-8858-760e77d45194/Screenshot_2020-08-05_at_12.56.01.png

Image Classification Template