SPEAR-Language-Packs

Language Packs for SPEAR-ASR and SPEAR-WakeUp are general-purpose language models that can be used for transcriptions, dictations and conversational AI.

If you need:

• to extend the existing language models
• other spoken languages such as Spanish
• domain specific language models such as for Medics
• limited vocabulary language models for Command and Control

contact us at [email protected].

For each Language, we provide a data folder and a model folder separately.

Data Folder

Each data folder should have two sub folders named as SPEAR-ASR and SPEAR-WakeUp. The content of these subfolders will be used to initialize SpearSdkApi. When using android API, you can put the data folder on any phone HDD and call the following codes to create SpearSdkApi, where LM_Package is the absolute path of the data folder.

spearSdkApi = SpearSdkApi.getInstance();
spearSdkApi.initialize(getApplicationContext(), LM_Package);

Model Folder

The model folder holds several fsts, some of which are used to define tasks for SPEAR-ASR, some of which are used to define tasks for SPEAR-WakeUp.

SPEAR-ASR

When using android API, you can put the corresponding fst on any phone HDD and call the following codes to create SpeechRecognizer, where example.fst needs to be replaced with the absolute path of the fst.

speechRecognizer = spearSdkApi.createSpeechRecognizer();
speechRecognizer.setFstGrammar("example.fst", true);

SPEAR-WakeUp

When using android API, you can put the corresponding fst on any phone HDD and call the following codes to create SpearWakeUp, where WAKEUP_FST needs to be replaced with the absolute path of the fst.

spearWakeUp = spearSdkApi.createSpearWakeUp();
spearWakeUp.initWithFst(WAKEUP_FST);