Detect problematic data slices in unstructured and structured data – fast.
Sliceguard helps you to quickly discover problematic data segments. It supports structured data as well as unstructured data like images, text or audio. Sliceguard generates an interactive report with just a few lines of code:
from sliceguard import SliceGuard
sg = SliceGuard()
issues = sg.find_issues(df, features=["image"])
sg.report()Install sliceguard by running pip install sliceguard[all].
Go straight to our quickstart examples for your use case:
- 🖼️ Unstructured Data (Images, Audio, Text) – 🕹️ Interactive Demo
- 📈 Structured Data (Numerical, Categorical Variables) – 🕹️ Interactive Demo
- 📊 Mixed Data (Contains Both) – 🕹️ Interactive Demo
We maintain a public roadmap so you can follow along the development of this library.