How to create jobs with pre-labeled data?
You can import your model predictions (or annotations created outside of GT Studio) and load them as editable annotations on a Job in GT Create. This can improve the labeling speed and reduce manual labeling effort.
Setup Annotation Specification
Add all the classes that are present in your annotation data.
Build Name
should match the annotation label in your pre-labeled data. If the build name is difficult to understand, you can set a differentDisplay Name
to keep it simple for the annotators.
Upload Pre-annotated jobs
Once you have set up the annotation specification, you can upload pre-labeled jobs to any workflow. If you have not created a workflow yet, please refer: Set up a Workflow
The annotation data needs to converted to GT Studio data format. You can refer to the data structure for the annotation data for the required base type and annotation type 👉
You can then put the job build with annotation data in the
Data
parameter of the Job Creation API defined below.
Create a job
POST
https://api.playment.io/v1/projects/:project_id/jobs
This endpoint allows you to create a job
Path Parameters
Name | Type | Description |
---|---|---|
project_id | string | ID of the project in which you want to create the job |
Headers
Name | Type | Description |
---|---|---|
x-api-key | string | API key for authentication. You can get it from the Settings section in GT Manage |
Request Body
Name | Type | Description |
---|---|---|
data | string | The data object contains the URLs for the images and the annotation data (if any). |
batch_id | string | A batch is a way to organize multiple jobs under one |
tag | string | A |
reference_id | string | The unique identifier of the job |
Editing Annotation in GT Create
After creating jobs with Annotated data, you can edit the annotations to refine the model predictions in GT Create.
To learn more about all the features in GT Create, refer:
Tools for LabellingLast updated