This page allows creating Dataset for Vision and Language Einstein. You can go to this page by clicking on Create link on Dataset tab. Also, you can see button name as Create Dataset on Dataset list page.
1. BASIC INFORMATION: Fill the name of Dataset in Name field.
2. API SETUP:
A. Dataset Type:
Image: When uploading image DataSet. (Available in v1)
Multi Label Image: When uploading multi label image. (Available in v1 and v2)
Intent: For upload text for intents. (Available in v1 and v2)
Sentiment: For upload sentiment DataSet. (Available in v1 and v2)
B. Way of Uploading:
URL: Url of your dataset. (Salesforce Recommend for this)
Labels: It comes if Dataset Type is Image or Multi Label Image (Try to avoid this to get better performance)
C. API Type:
Asynchronously: Entry of DataSet get created but uploading runs in background (Use when file is more than 10 MB)
Synchronously: Dataset get created instantly (Use when file is less than 10 MB)
I recommended you to use Asynchronously. Because you can have Dataset more than 10 MB as well. So you don’t need to check about Dataset size.
3. CONTENT TO UPLOAD: You define here URL or Labels.
For the Image and Multi Label Image, you must give URL of the .zip file
For the Intent and Sentiment you must give URL of the .csv, .tsv , .json
You will get Labels input box if you have selected Labels in Way of Uploading option.
Note: When specifying the URL for a .zip, .tsv, .csv., .json file in a cloud drive service like Dropbox, be sure it’s a link to the file and not a link to the interactive download page. For example, I uploaded my Dataset to GitHub
You should not use URL like above. It must be a direct link to file like:
Now click on Save button. You will be redirected to Dataset detail page if all things are good without error.