I ran into the same error while finetuning a glow-tts model, and I had to update the config.json file. In the datasets parameter, I added:
config.json
datasets
"formatter": "ljspeech",