In Chapter 5, Salesforce Einstein Vision, we trained the dataset using CURL. We zipped all the images in our local drive and then used the CURL command to create the dataset from the .zip file asynchronously.
Currently, Einstein supports the creation of a dataset via three major APIs:
- Creating a dataset from a zip file synchronously
- Creating a dataset from a zip file asynchronously
- Creating a new dataset and labels, if they're specified
To derive better and more efficient performance, the docs refer to creating dataset from a zip file asynchronously, and hence for rest of the chapter, as a recommended practice, we will be using asynchronous creation of dataset from the zip file.
For data storage, we will leverage the free version of Cloudinary (https://cloudinary.com/). The primary reason to store the files used for training in a cloud...