XFUND: A Multilingual Form Understanding Benchmark
Introduction
XFUND is a multilingual form understanding benchmark dataset that includes human-labeled forms with key-value pairs in 7 languages (Chinese, Japanese, Spanish, French, Italian, German, Portuguese).
Three sampled forms from the XFUND benchmark dataset (Chinese and Italian), where red denotes the headers, green denotes the keys and blue denotes… See the full description on the dataset page: https://huggingface.co/datasets/FrancophonIA/XFUND.
