Local dataset quality check
Check class imbalance in an image dataset
One class can dominate a dataset while important rare classes have too few examples in a split.
What to check
Compare per-class counts and split coverage before deciding whether to collect, relabel, or rebalance data.
Why it matters
An imbalanced dataset can produce deceptively strong aggregate metrics while under-serving rare classes.
When Pro helps
Pro supports larger datasets and repeatable quality checks.
What your report can look like
Class coverage review
An illustrative example of the findings and repair guidance Labelyst produces for your dataset.
Status: Ready with warnings
Score: 76/100
2 findings
Findings
- 1 class has limited training coverage
- 1 split is missing a rare class
Passed check
- Class names are consistently formatted
Suggested next step: Collect or deliberately document the rare-class gap.
Rules: standard profile · local check
Illustrative local result. It is not a certification.
Check it before training
No account or upload is required for a Free check. Analysis, repairs, and reports stay on your device.