batchTranscribe runs minimal recognizers.

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

Closed KatieProchilo closed 5 years ago

KatieProchilo commented 5 years ago

Total number of recognizers is computed and used in batchTranscription()'s for loop rather than the number of concurrent calls.
Renamed variables local to batchTranscription().
Use duplicate code only once.

HobbyProjects commented 5 years ago

This code path has no tests. Feel free to add some to cover your changes if you can.

KatieProchilo commented 5 years ago

@HobbyProjects I'll create an issue. At a glance tests seem non-trivial.