Important Note
Please don't forget to check the version history of LMS as it is a system that allows users to store speech recognition models.
ENHANCEMENTS
Common Changes
- Support nbest result for e2e models
- Removed License Service setup, SR setup will not install License Service.
REST Changes
- Added full http address type.
MRCP Changes
- Added mrcp/schemes http services.
IMPROVEMENTS
- Fixed confidence computation for http recognitions.
KNOWN ISSUES
- [ERROR] SestekRecognizer : lib-fstbuilderk : Locale is not set properly" error is written in the log in Persian and Pashto grammar and mrcp recognitions.
Technical Details (On-Premise Only)
Build Number: v11.4.0
LMS - Language Model Service
v11.4.0
ENHANCEMENTS
- Changed web tabbed layout to single page.
- Added secondary service host. For now secondary service supports only getLatestVersionsOfAllVariantsOfAllModels end point.
- Do not show error dialog if overwrite model file operation is canceled by user
IMPROVEMENTS
- Updated onnxruntime to v1.14.1.
KNOWN ISSUES
- If a Turkish character header is added in the header "CustomResponseHeaders" settings in the LMS AppSettings.json file, "Invalid non-ASCII or control character in header" is written in the LMS logs. In this case, the requested headers will not be added.
- Pronunciation of the word cannot be added to the Arabic language model via LMS Explorer.
- There is no healthcheck method in this version.