How to interpret the results of the Japanese document?

I am trying to convert a Japanese document into text chunks as mentioned in 'Financial Report Chunking for Effective Retrieval-Augmented Generation.' Can you tell me how to use the Unstructured Inference API to get the output?