The LangIdOptimize
task optimizes the balance between language classifiers. After training, some classifiers might be stronger than others because of properties of the training material and the languages in question. The optimization process weights the language models so that weaker languages have increased accuracy, without compromising accuracy for stronger language models. This process improves consistent performance.
Parameter | Description | Required |
---|---|---|
Type | The task type. Set to LangIdOptimize . |
Yes |
ChunkSize | The size in seconds of each chunk of audio. | |
ClassList | A list of language classifiers. | |
ClassPath | The path to the directory containing the language classifiers. | |
DataList | A list of language feature files. | Yes |
DataPath | The path to the directory containing the language feature files listed in the DataList. | Yes |
LangList | A subset of languages to use from the classifier list file. | |
Log | The name of the log file to write. | |
Out | The classifier list to write the optimized weights to. | Yes |
The ClassList
parameter is required only if you want to change the audio sample rate, or if you want to use your own custom classifiers. You might also need to specify the ClassPath
parameter, depending on the location of the classifier files.
http://localhost:13000/action=AddTask&Type=LangIdOptimize&DataList=ListManager/OptList&DataPath=C:\LangID\Data&ClassList=ListManager/ClassList&ClassPath=C:\LangID\class&Out=ListManager/OptClassSet
This action uses port 13000
to instruct HPE IDOL Speech Server, which is located on the local machine, to test the feature files listed in the OptList
list against the classifiers specified in the ClassList
list and then generate an updated classifier list–OptClassSet
.
The LangIdOptimize
action generates a log file containing information about the optimization process. For details about this log file, see the HPE IDOL Speech Server Administration Guide.
|