Do you want to request a feature, report a bug, or ask a question about wit?
Report a bug.
What is the current behavior?
This applies to the language dutch (NL).
I am using the HTTP speech endpoint to transcribe audio and to extract a number using the wit/number entitiy.
For the numbers 100 - 999 the transcribed text is usually like the following:
Honderddrieënzestig: <-- this is not recognized as a number, and wit.ai does not extract a number (here it should be 163)
Sometimes the transcribed text looks like:
Honderd tweeënzestig: <-- here the number will be correctly extracted (here it is 162)
So, in conclusion, if there is not a space between "honderd" and the rest of the number, the recognition of the number fails.
Examples where the extraction of the number fails:
Do you want to request a feature, report a bug, or ask a question about wit? Report a bug.
What is the current behavior?
This applies to the language dutch (NL).
I am using the HTTP speech endpoint to transcribe audio and to extract a number using the wit/number entitiy. For the numbers 100 - 999 the transcribed text is usually like the following:
Honderddrieënzestig: <-- this is not recognized as a number, and wit.ai does not extract a number (here it should be 163)
Sometimes the transcribed text looks like: Honderd tweeënzestig: <-- here the number will be correctly extracted (here it is 162)
So, in conclusion, if there is not a space between "honderd" and the rest of the number, the recognition of the number fails.
Examples where the extraction of the number fails:
If applicable, what is the App ID where you are experiencing this issue? If you do not provide this, we cannot help. 612490360532442