Joll59 / d-ser-t

d-ser-t quantifies speech recognition accuracy of the MSFT speech service and/or user created MSFT custom speech service models.
2 stars 2 forks source link

Nipoulad/falsepositive #52

Closed HobbyProjects closed 5 years ago

HobbyProjects commented 5 years ago

Add the ability to consume an exception list where a list of known transcriptions can be replaced with correct replacements. Also, allow the specifications of regex patterns so they can be added via the list without having to hardcode them in the code. All above comes with tests to verify.