When training the Marylux-648 dataset with the embedded gruut-phonemizer of the Coqui-TTS Glow-TTS model, the audio- and text-data are not aligning. After debugging the problem, I discovered that the reason is a wrong phonemisation because the required files are not stored in the correct folder, due to a wrong setup of the project.
As a workaround, I moved the related files in the specified folder and restarted the training. Now the alignment works as expected.
When training the Marylux-648 dataset with the embedded gruut-phonemizer of the Coqui-TTS Glow-TTS model, the audio- and text-data are not aligning. After debugging the problem, I discovered that the reason is a wrong phonemisation because the required files are not stored in the correct folder, due to a wrong setup of the project.
As a workaround, I moved the related files in the specified folder and restarted the training. Now the alignment works as expected.