We’re excited to announce our new Python SDK for Labelbox, launching today in beta.
While you can interact with Labelbox through our powerful GraphQL API, we also recognize that Python is the most common programming language used by data scientists and the machine learning community at large. In order to enable easier workflows for those most comfortable with Python, we set out to create a pip package that you can install and use out of the box. You’ll be able to avoid writing and maintaining your own scripts, and writing queries and mutations that fit within our schema, without ever having to learn GraphQL.
With the Python SDK, you can now:
- Simplify your data import
We’ve simplified the data import process for Labelbox so that you can use bulk DataRow creation. Best yet, this process is asynchronous, meaning you can (but don't have to) wait for bulk creation to finish before continuing with other tasks (or even terminating the client).
- Interact with the API in an object-oriented way
Create projects and datasets programmatically, export labels, and add metadata to your assets all in an object-oriented way—complete with all relationships between objects.
- Set queue customization
With queue customization you can support active learning by adjusting your labeling queue to focus on improving confidence in specific classes. We also see customers use queue customization to support time-sensitive labeling, such as real-time labeling workflows that are part of human-in-the-loop AI products and services.
We’d love for you to participate in the open beta and provide feedback. There's no need to signup, you can access the Python SDK directly today via
pip. You can always leave feedback in Labelbox via Intercom chat box where we have teammates at the ready.
pip install labelbox
You can also learn more about the Python SDK beta and our GraphQL API in our docs.