I used the Supper at Emmaus by Caravaggio as a training test to see if it would hold up with only one image.

I used my cropping tool to separate it into 15 smaller images to use as data to train a LoRa. My image cropping tool can be found at: https://image-bucket.com/

https://en.wikipedia.org/wiki/Supper_at_Emmaus_(Caravaggio,_London)#/media/File:1602-3_Caravaggio,Supper_at_Emmaus_National_Gallery,_London.jpg