How to handle the error throw from dgraph bulk

I had a large json file which encoded wiht utf-8 to init my Dgraph app with dgraph bulk.
while It throw error: invalid character '\b' in string literal and I was sure the character not in my file.
I know there may be some character coding problem which cause the problem beacuse lots of Chinese character in my file.
How can I find which predicate cause the error? Is there any way to just ignore the error to make dgraph bulk goes on?

you can try this flag

dgraph bulk -h | grep ignore
      --ignore_errors                ignore line parsing errors in rdf files


gzip -t

-t --test
       Test. Check the compressed file integrity.|+dgraph