"Include each sentence of the corpus on its own line, and terminate each line with a carriage return. Including multiple sentences on the same line can degrade accuracy."
The data parser for all the txt files should ensure that the resulting output has each sentence on a new line.
The Watson STT documentation specifies:
"Include each sentence of the corpus on its own line, and terminate each line with a carriage return. Including multiple sentences on the same line can degrade accuracy."
The data parser for all the txt files should ensure that the resulting output has each sentence on a new line.
https://cloud.ibm.com/docs/services/speech-to-text?topic=speech-to-text-corporaWords#prepareCorpus