In this chapter's project, our assets will primarily be the training images we have collected and, indirectly perhaps, classified. These image assets are added in a way similar to the process we used to add data assets to Watson Studio projects in earlier chapters, but there are a few differences, which we will soon see.
For now, we will perform the following steps:
- Go to the Assets tab and, under Models, click on New Visual Recognition model:
- Once the model is created (it should take only a few moments), you can browse to or drag and drop the training (.zip) files we collected in the earlier section of this chapter to add them to our new project. This will upload the image files to Cloud Object Storage (COO), making them available to be used in our project:
You do not have to load each image file independently, just the three zipped files (happy.zip, sad.zip, and negative.zip). The ZIP files should then be listed as shown in the preceding screenshot.
Although you can upload the ZIP files, Watson will not allow those ZIP files to use the Preview feature on the Data Assets page. This is not a problem though, as you can still preview the images from the Model page, as we will see shortly.