You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, I'm using DeepSpeech 0.9.3 and trying to find out Word Error Rate on the Common Voice Mozilla dataset. My question is, is there any python script to find out WER? if Yes, please help me with this.
The text was updated successfully, but these errors were encountered:
@Chaitanya1597 and @yunzqq - what you are looking for is the metric called - "Levenshtein Distance". You have to calculate this between the Ground-Truth-Text (alias Referenztext (the origin voice input that you made as text)) and the Deepspeech - Transcription (output text)
more useful informations: Articel from Rafael C. Carrasco - i hope this is helpful for you?
Currently, I'm using DeepSpeech 0.9.3 and trying to find out Word Error Rate on the Common Voice Mozilla dataset. My question is, is there any python script to find out WER? if Yes, please help me with this.
The text was updated successfully, but these errors were encountered: