Update README.md

README.md

@@ -72,6 +72,8 @@ Training data comes from: [ENC-PSL/odil-medieval-folio-illumination-bin-dataset]
 
 ## Distribution of data
 
+## Distribution of data
+
 - illuminated
   - train: 519
   - dev  : 112
@@ -81,6 +83,10 @@ Training data comes from: [ENC-PSL/odil-medieval-folio-illumination-bin-dataset]
   - dev  : 111
   - test : 112
 
+> **Data augmentation.** During training, data augmentation was applied to the training split only in order to improve robustness and reduce overfitting.
+>  The augmentation pipeline included random horizontal flips, small random rotations up to 5°, and light color jittering with brightness `0.12`, contrast `0.12`, saturation `0.08`, and hue `0.02`.
+> Validation and test images were evaluated without augmentation.
+
 # What counts as "illuminated"?
 
 ### Positive (illuminated)