Skip to content

instadeepai/SKAInnotate

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

27 Commits
 
 
 
 
 
 
 
 

Repository files navigation

SKAInnotate

SKAInnotate logo by Kobby

SKAInnotate is a data annotation tool designed for Google SKAI and operates within Google Colab Notebooks. It consists of two main notebooks: one for project administration and another for annotators.

Project Admin Notebook (Admin-Notebook.ipynb)

The Project Admin notebook provides the following capabilities:

  1. Create Annotation Projects: Set up new annotation projects.
  2. Manage Annotators: Add or remove annotators from projects.
  3. Configure Project Settings: Define project titles, storage buckets, maximum annotators per task, and other configurations.
  4. Assign Tasks: Assign annotation tasks to annotators.
  5. View and Export Annotations: Access and export annotations from all annotators.

Setup

Prerequisites

To use SKAInnotate effectively, ensure the following prerequisites are met:

  • Google Cloud Project Permissions: Configure the Google Cloud project with permissions to manage users and resources. The following roles should be sufficient for the admin; Cloud SQL Admin, Storage Admin and roles/resourcemanager.projectIamAdmin.
  • Data Upload: All data to be labeled must be uploaded to a Google Cloud Storage bucket.

Adding Data for Labeling

To add data for labeling, follow these steps:

  1. Place a metadata CSV file in the same path as the data.
  2. The CSV file must include the following columns:
    • example_id: Unique ID for each example.
    • image: Unique filename for each example.

Global Configurations

Configure the following global settings for your SKAInnotate project:

  • project_id: Google Cloud Platform (GCP) project ID.
  • region: Location of the GCP project.
  • instance_name: Name of the Google Cloud SQL instance.
  • root_password: Password for the PostgreSQL database user.

Project Configurations

Define project-specific settings:

  • project_title: Title of the project.
  • bucket_name: Name of the Google Cloud Storage bucket.
  • bucket_prefix: Prefix for the bucket (e.g., my-storage/inner-storage for data path gs://my-bucket/my-storage/inner-storage/image1.png).
  • comma_separated_labels: Labels stored in the database as a string. Include "skip" to allow annotators to skip uncertain labels.
  • max_annotators_per_example: Maximum number of annotators per task to provide different perspectives.
  • completion_deadline: Deadline to complete annotation tasks.

Annotator Notebook (Annotator-Notebook.ipynb)

The Annotator notebook is used by annotators to perform the following tasks:

  1. Receive Assigned Tasks: Retrieve tasks assigned by the project admin.
  2. Label Tasks: Annotate data using a user interface.
  3. Write Annotations: Store annotations securely in a central cloud database.

Setup

Annotators need to set up the following configurations:

  • project_id: GCP project ID.
  • region: GCP project location.
  • instance_name: Google Cloud SQL instance name.
  • Authentication: Annotators require a username and password provided by the project admin to access assigned tasks.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published