image recognition with appinventor

41 views

Skip to first unread message

Amira Ibrahim

unread,

Oct 18, 2021, 6:57:17 AM10/18/21

to Machine Learning for Kids

Dear Dale,
I'm confused, hope you could me help me.

In image recognition projects with app inventor, why should I use the train new model block before sending an image to be classified?

Doesn't the model suppose to be ready?

As I trained before on machine learning for kids.

why should I train the model each time I'm using the app?

Each time I open the app, I have to go through these steps:

1-training button

then

2-checking the status

then

3-capture the photo

Dale Lane

unread,

Oct 18, 2021, 3:08:13 PM10/18/21

to Machine Learning for Kids

It depends on whether you are training your model on your computer (or in this case, on your Android device) or in the cloud.

If the model is in the cloud, you are correct - training it on the web means it will be available for use on your App Inventor device.

If the model is on your device, then you will need to train it both on the web (creating a model on your computer/laptop) and on your App Inventor device (creating a model on your mobile).

This is described in more detail in What is the difference between training image models "on your computer" or "in the cloud"? in https://machinelearningforkids.co.uk/help