Get up and running with Watchful by following the steps below.
## Login aws ecr get-login-password --region us-west-1 | \ docker login \ --username AWS \ --password-stdin \ 610410161133.dkr.ecr.us-west-1.amazonaws.com ## Run the Container docker run \ --name watchful \ -d -p 9001:9001 \ -v ~/watchful:/root/watchful \ -e WATCHFUL_LOG=info \ 610410161133.dkr.ecr.us-west-1.amazonaws.com/production/watchful:3.0.1234
You can access the import modal by clicking
Dataset in the menu bar.
Once data has been imported, create the classes you want to start labeling for by, going to the input box, just right of the words "Selected Class", writing out the name of the class, and hitting enter. If you're not sure what your classes are at this point, you can always create them one-by-one as you go.
Class names must be of the following format:
- Class names must start with an uppercase letter
- Can only contain alphanumerical characters or '_'
- Cannot exceed 32 characters.
Watchful has two class options NER Named Entity Recognition or FTC Full Text Classification. Class selection affects some query features and UI modes that are class specific.
Select a class to start labeling for by simply typing its name in the "Selected Class" input box.
Click into the "Hand Label" tab in the center of the nav bar, and label a few pages of candidates until the Base Rate chart stabilizes. This will help give Watchful some initial signal to calibrate it to your class.
To learn more about Hinter Suggestion types read more on Hinters
To learn more about expectation best practices see Hinters
Go to the settings page, give your self a username, and connect to your organization's Watchful Hub. More on Setting up Watchful Hub here.
Share your project
Share your progress and incorporate other's by going into the collaboration view. Pull in other team member's actions from Watchful-Hub, and Push your own actions.
Updated 6 days ago