Phoneme input error

[JeneeeF]

Hi,
Thanks for your great work.
I ran into an issue with the Native mode when inputting Chinese Pinyin with tone numbers, for instance: "这几日 {zuo1} 坊里多了个小客人 —— 隔壁搬来的小女孩念念。小姑娘扎着羊角辫，总踮着脚扒着 {zuo4} 坊门槛", the synthesized audio pronounces the tone numbers (1 and 4) as literal digits instead of applying the corresponding tones to the syllable "zuo".
I’m wondering if yout team has tested this scenario before, and if there’s a way to avoid this problem?

meiyangyang_output (1).wav

[JeneeeF]

I know how to avoid the synthesized audio pronounces the tone numbers as literal digits, just turn off the text normalization. But another phenomenon appeared: when I specify the correct pinyin for a polyphonic word, the model output is pronounced in a different wrong way. For instance, the input text is: "蛋糕公司马上{diao4}配人手来着手处理价格上{tiao2}的舆情。"
Is there a mismatch between pinyin and actual pronunciation in your training data?

meiyangyang_output2 (1).wav