AWS sagemaker documentation change
Summary
Removed steps 3-4 from data validation process (domain expert review and statistical analysis), simplified quality validation procedure
Security assessment
Changes affect data validation documentation but don't address security aspects. The removed steps relate to general data quality rather than security controls.
Diff
diff --git a/sagemaker/latest/dg/nova-distillation.md b/sagemaker/latest/dg/nova-distillation.md index 5596ad9c6..5b1bd0363 100644 --- a//sagemaker/latest/dg/nova-distillation.md +++ b//sagemaker/latest/dg/nova-distillation.md @@ -426,5 +426 @@ Before proceeding to fine-tuning, it's crucial to validate the quality of the au - 3. If possible, have domain experts review a larger subset of the augmented data. - - 4. Consider running basic statistical analyses on the full dataset to check for unexpected patterns or anomalies. - - 5. If the quality doesn't meet your expectations, you may need to adjust your input data or distillation parameters and rerun the data augmentation process. + 3. If the quality doesn't meet your expectations, you may need to adjust your input data or distillation parameters and rerun the data augmentation process.