Did I make the wrong choice? and my options going forward.

[Dr. James Burke]

Background: I'm a high school CS teacher with a background in software and STEM ed research, and I've seen how great Machine Learning for Kids is. I've ordered the book and hope to integrate some ML into an introductory CS course that's built around the idea of making games.

I'm playing with Machine Learning for Kids in my classroom now, since I have some flexibility at the end of the year, and I created a class of 30 students. I've spread the student logins over several real classrooms with the intention of having students do group work in which only one student in the group logs in (as the "driver"). So, I didn't intend to need more logins.

I chose an unmanaged class and didn't realize I'd have a limit of 5 simultaneous models. So it seems I may have messed that up, as one of my classes has 8 groups in it; obviously they can't all have models at the same time. 

Is there an option for more than 5 models at a time outside of the paying IBM? I'm not averse to exploring talking to IBM if necessary, but I would like to know if I've unnecessarily made this harder on myself by choosing an unmanaged account. I had a problem with Try it Now earlier, but I'm willing to give that way of working a shot if it's the intended way to use Machine Learning for Kids.

Cheers,

-JP

[Dale Lane]

Hi

I already replied to your other post before seeing this, so I'll start by copying what I wrote there and extending it

An unmanaged class account doesn't mean you can only create five text models at a time (and there are in fact many thousands of unmanaged class accounts with many more than five text models).

Unmanaged means you provide your own Watson API keys.

Managed means I sort out the Watson API keys for you.

That is the only difference. 

There are no other differences about limits on the number of models. I haven't written a line of code to try to limit the number of models created in a class. I haven't even written code to count the number of models you have in your class. 

But if you choose to use your own Watson API keys instead of using mine, then this does mean you need to provide enough API keys to meet the needs for your class. 

And if you don't do that, you'll see errors - errors from the IBM Watson API that the Machine Learning for Kids website is just passing back to you. 

This includes if you provide a Watson API key and then delete your IBM Cloud account. That would stop your API key from working, and I'll return the error I get from IBM Cloud when you try and train an ML model with it. 

Similarly, there are free ("Lite") plans available from Watson that can only be used to create five models at once. So if you try to use it to train more than five ML models, that won't work, and I'll return the error I get from IBM Cloud when you try. 

Any errors like these are attributes of the API key you have created and provided, not something that I am enforcing. It's just that IBM Watson has usage limits on it's API keys, and they won't allow you to exceed those limits.

> Is there an option for more than 5 models at a time 

> outside of the paying IBM?

Yes, managed class accounts. Ask me to set up the API keys for you. I don't use "Lite" plan API keys for this.

Would you like me to change your account to be a managed class account?

Kind regards

D

[Dr. James Burke]

Dale,

> Would you like me to change your account to be a managed class account?

Yes, that would be wonderful. Thank you!

-JP

[Dale Lane]

I've done that now

Please also note that the 30-students limit is a default, not a hard limit. If you need that to be increased - please let me know.

Kind regards

D