- Now that you have a project, the next step is to create an experiment. Go back to create a new experiment and fill in the details.
-
In order for you to create an instance, you will need to create a Watson Machine Learning instance as well. IBM WML is a comprehenive solution that is your gateway to various machine and deep learning technologies.
-
If you have existing Waton Machine Learning instance then you should be able to click on the
Reloadbutton and associate that account and skip step 4.
4a. If not, click on the Associate a Machine Learning service instance link and follow the steps to create an instance.
4b. Select the Lite plan and click on create. Once created close the new tab that had opened up and come back to the experiment assistant
- You should be able to click on the
Reloadbutton for the Machine Learning instance and your instance for Machine Learning should show up for selection.
-
Next step is to hook the cloud object storage instance that you had created in the last step. Click on
Selectbutton forCloud Object Storage bucket for storing training source and results filesand you will be redirected to a page asking you to connect your cloud object storage account with the experiment. -
Click onthe tab that says
New connectionand select the cloud object storage instance that you had created earlier.
-
You will need a training/data bucket to hold your training data and a results bucket to hold your results.
-
Select the
Newradio button for creatingBucket containing training dataandBucket for storing training results.
** NOTE: Bucket name is restricted to lowercase letters from a to z, numbers, or dashes, between 3 and 64 characters in length.**
- The bucket names are globally unique, so be careful in picking up names for your train/results bucket. You can add a suffix of your username to make them unique.
- Next step it to go ahead and create a training definition which will be used to specify details about your training. Follow the instructions here