Train my own annotated data to use Funsd format

[xia-zhenjun-tfs]

I tried to process my data in your Nielsr/funsd-LayoutLMv3 format, but there was a coding error

Extracting data files: 100%|██████████| 1/1 [00:00<00:00, 1.58it/s] 51 tables [00:00, 169.73 tables/s]Failed to read file 'C:\Users\zhenjun.xia.cache\huggingface\datasets\downloads\extracted\3fa853741480b9539086141ffbea32fab1262caeafb97e5e6e5af3ce15d35298\dataset\testing_data\images\1595519327454_0.png' with error <class 'pyarrow.lib.ArrowInvalid'>: JSON parse error: Invalid value. in row 0 Traceback (most recent call last):

'utf-8' codec can't decode byte 0x89 in position 0: invalid start byte

How should I solve this error, thank you very much