Failed to phonemize: ./training/{voice}//valle/*.phn.txt

[ffd000]

Vall-e doesn't create the .phn.txt files when preparing my dataset, but it does create the train.txt file with the transcript and the .qnt.pt files.
I have 6 vocal clips each 10 minutes long. These are automatically segmented when I prepare the dataset and transcribed, but it says "failed to phenomize" on every one, so no .phn.txt files are created, and these are necessary for training as far as I understand.
python -m vall_e.emb.g2p training/{voice} does nothing.
I'm using https://git.ecker.tech/mrq/ai-voice-cloning webui with vall-e module.