Learn about Galileo’s Data Error Potential (DEP) score, a metric to identify and categorize machine learning data errors, enhancing data quality and model performance.
Galileo Platform surfaces mislabeled, garbage samples by ordering in desc order of DEP score
Galileo DEP score | Model confidence score |
---|---|
![]() Noise recall: 99.2% | ![]() Noise recall: 87.5% |
![]() Noise recall: 89.0% | ![]() Noise recall: 64.9% |
Benchmark (Train): Performance on Noisy Datasets
Benchmark (Train): Final Epoch Train Performance
Benchmark (Train): Average across all epochs
Benchmark (Test): Average across all epochs