In project we are using speech-to-text google service.
It brings a lot of advantages and simplifies audio processing on one hand, on other hand we heed special data formatting for example:
Field Type : Money
Current Output : no payment due
Expected Output : $0
Current Output : no outstanding balance
Expected Output : $0
Field Type : Date
Current Output : May 1st 2020
Expected Output : 05/01/2020
Current Output : May 1st 2020.
Expected Output : 05/01/2020
Current Output : June 1st 2020.
Expected Output : 06/01/2020
Current Output : March 13th.
Expected Output : 03/13/2020 (automatically infer the year)
We've tried to look at class-token, but this approach did not provide tool to cover our needs. We've checked the API few times, researching behavior - result still the same. I believe that we've missed something and hope for community help.
P.S.
Our project language is java and we believe that in scope of the issue it is not so important.