Streamline Data Annotation in Jupyter Notebooks with Pigeon and ZenML
Integrate Pigeon, a lightweight and intuitive data annotation tool, with ZenML to effortlessly label your datasets directly within Jupyter notebooks. This integration simplifies the annotation process for text classification, image classification, and text captioning tasks, making it ideal for quick labeling during the exploratory phase of your ML projects.
Features with ZenML
- Seamless Integration with Jupyter Notebooks
Annotate your data without leaving your familiar Jupyter notebook environment, ensuring a smooth workflow. - Easy Setup and Configuration
Installing and registering the Pigeon annotator with ZenML is a straightforward process, requiring minimal effort. - Efficient Data Management
Utilize ZenML's annotator dataset commands to easily list, delete, and retrieve statistics for your annotated datasets. - Streamlined ML Workflows
Incorporate Pigeon annotations seamlessly into your ZenML pipelines, enabling efficient data labeling within your ML workflows.
Main Features
- Ultra-lightweight and open-source
- Supports text classification, image classification, and text captioning
- Intuitive interface for quick and easy labeling
- Ideal for small to medium-sized datasets
- Facilitates collaborative labeling within Jupyter notebooks
How to use ZenML with
Pigeon
from zenml.client import Client
annotator = Client().active_stack.annotator
annotations = annotator.launch(
data=[
'This movie was fantastic!',
'I was disappointed by the ending of the book.'
],
options=[
'positive',
'negative'
]
)
The code example demonstrates how to use the Pigeon annotator within a Jupyter notebook using ZenML. It launches the annotator with a list of text data and predefined label options. The annotator returns the labeled data as a list of tuples, each containing the text and its corresponding label.
Additional Resources
View the Pigeon GitHub repository
Read the full Pigeon integration documentation
Explore the Pigeon Python package on PyPI
