A multilingual image editing instruction dataset derived from Apple's Pico-Banana-400K, containing 12,424 triplets (source image, edited image, instruction) with instructions translated into 4 languages: English, Nepali, Bengali, and Hindi.
images/
source/ # 12,424 source images (PNG, 512px)
target/ # 12,424 edited images (PNG, 512px)
metadata.jsonl # Full metadata with translations
Fields in… See the full description on the dataset page: https://huggingface.co/datasets/Legend2727/xLingual-picobanana-12k. 
