In this tutorial, you'll set up an IBM Watson account, then train a custom visual recognition Watson service model, and set up an iOS app to use the exported Core ML model.
Do you happen to know the model that IBM is using for the image recognition? Squeeze Net or other?
sorry, no I don’t — by rights, it should be some whizbang Watson thing? Get on the Swift@IBM Slack and ask 
Here’s a direct link to the Swift@IBM Slack. Feel free to ask specific questions in the #apple-development channel 
I followed the steps mentioned in the tutorial but after clicking on Create Model, and filling in the project name and tap on Create, I am not getting the Default Custom Model under the created project.
It seemed initially the UI has changed. So I uploaded the assets as shown in the screenshot but I am not getting option to Add selected to model and also not getting Train Model.
I checked the Github page too (GitHub - watson-developer-cloud/visual-recognition-coreml: Classify images offline using Watson Visual Recognition and Core ML). But there also has been no change in the steps.
Can anyone kindly help me where am I going wrong? Thanks.
hi Sandipan: I’ll check this out right away! there are some IBM people here at RWDevCon who might be able to help.
stay tuned …
Hi sandipanb,
Looks like you are in the project view. Click on Assets, look for the Visual recognition models section, then click on the model you are trying to train. From there you should be able to add data.
Hi devinaconley,
I am encountering the same issue as sandipanb. I can’t see a way to add a model in the Assets tab.
Under Assets, all I see are
Data assets (w/o any option to add a data asset)
Streams flows (w/ option to add a new stream flow)
Dashboard (w/ option to add a new dashboard)
Data flows (w/o any option to add a data flow)
I’ll dig around, but wondering if there is a higher-level profile setting that needs to be enabled in order for us to add the data model.
If you have any insights, please let us know,
Thanks
Kartik
Hi ksubram1,
Go to the Settings tab. Under the Tools section, make sure that Visual Recognition is checked. Also, under Associated service, make sure that you have associated an instance of Watson Visual Recognition.
Hi devinaconley,
Thanks for your response and help! Unfortunately, in Settings->Tools, I don’t see Visual Recognition. I did associate an instance of Watson Visual Recognition with the Model under Associated service . See screenshots below. I’ve tried to delete the project and start from scratch, but seem to run into the exact same situation everytime.
I have the same results as Kartik, i.e. no default model, apparently no way to create one, no Visual Recognition checkbox. I too deleted the project and got the same results.
Ok - I think I got it to work and was able to replicate Audrey’s default custom model setup.
Here is what I did:
Assumptions: You should have logged into your IBM Cloud account.
If so,
Go to IBM Cloud (this is your IBM Cloud Dashboard).
Then, click on the hamburger menu icon (the three dashes on the upper right side).
Click on Apple Development
Click on Watson Visual Recognition for Core ML (Get Starter Kit)
On this page, click on “Create Project”
Enter all your project details (Make sure you choose an organization that gives you the option to choose the WatsonDataPlatform space)
Hit “Create Project”
At this point, your project is created and you need to add the Visual Recognition service to your project. To do this, click on the “Add Resource” icon on the upper right and scroll down to select Visual Recognition and add that to your project.
Once you do this, then you should see “Visual Recognition” listed as a resource. Click on the “Launch tool” link
Here, you will see the option to create a Custom Model for Visual Recognition. Click “Create Model”
This will bring up the EXACT same page that Audrey had in her description and you should be able to add the assets and train the model on the same.
Hope this helps!
Kartik,
Thank you very much for the effort you put into this. Unfortunately it didn’t change the result for me. I attempted to follow your instructions twice using two different browsers but ended up in the same create project screen that I did following the author’s instructions. There were two differences though. At your step 6, I had no options other than the name which was pre-filled by the site. At your step 8, there was no Visual Recognition link. There was a Watson box and selecting it brought up a different screen that did have Visual Recognition so I selected that. Step 9 and 10 work as described but clicking Create Model takes me down the same rabbit hole as the original attempt at a tutorial.
Thanks again,
Mike
hi Mike: sorry you’re having so much trouble! did you get the _us-south appended to your login email, for the IBM Cloud account Organization? In the IBM Watson page, click the user icon in the upper right corner, select Settings, and check?
Although I can’t find any way to edit Organization to add _us-south if it’s missing. All the Watson services are available only in certain regions; if you’re not in one of them, appending _us-south gets you access. Update: at step 6 of @ksubram1’s reply above, you can select US South as the region to deploy in and that will change the organization email.
If that’s not the problem for you, I’ve asked Devin if there’s a site map to show how all these pages are connected. That might be somewhere to start …
thanks! yes, there are multiple paths to the same place, but I tried to show readers what looked like the most direct path. Starting from the cloud dashboard seems to have a lot more pathways leading out, but it also seemed much easier to get lost.
what I found disorienting was when the back button would take you someplace completely different…
Audrey,
I may be the outlier then. I missed the us-south requirement. At Kartik’s step 6 he mentioned using the right settings but I didn’t get an option. At login the site wanted to use my existing IBM account which is at least 10 years old and possibly from last century. I’ll try creating a new account and see if that changes things.
Thanks!
Mike
ah, yes, that region thing seems to be crucial! I’ll edit the tutorial to make it stand out more … and add that if you don’t see it, you might have an old-account issue etc.
thanks!
Well I finally made it to the Default Custom Model page but am a little hazy about how I arrived there since I had given up and was just clicking on pretty much anything. What I do remember is:
I created a new email address so I could create a new IBM account. I attempted to follow the tutorial. When I got to the Create Organization page it did not have the magical us-south and I could not add it. There was a button for editing though and I clicked that and that added the magic string. However that still led me down the same rabbit hole as my other attempts. I logged out and clicked something on the IBM home page and somehow arrived at the correct place. When I get finished chewing on my keyboard I could attempt to recreate the process if anyone else has similar issues.
I did notice that my original account did have US South as the region but it was not on the organization string.
Thank you,
Mike
now that you’ve logged into Watson, you should be able to click the bottom link (Already signed up? Log in) on the Watson studio login page from now on, and that should take you to the right place.
LOL. I’m just not destined to complete this. I got to the page where you upload images but they don’t seem to load. I let it run for 20 minutes the first time and 30 the second. I’m trying them one at a time now. 
And I do have decent broadband service
