No textual-like characters and punctuation such as {,},<,>,[,] will usually result in low-quality speech generated by the model.
No textual-like characters and punctuation such as {,},<,>,[,] will usually result in low-quality speech generated by the model.