chunked/nltk_chunking/tasks_text-to-speech/chunk_30.txt

To identify unsupported tokens, extract all unique characters in the dataset using the SpeechT5Tokenizer which 
works with characters as tokens.