no punctuation marks when transcribe Chinese.

[SteveTanggithub]

model:whisper-large-v3-turbo-v1.0.0.bin
we use the model to transcribe chinese speech, but there is no any punctuation marks in the output. And we only get the comma marks after we add config like "--prompt "punctuation." "
How can I get the correct punctuation marks for chinese?

[mrfragger]

--prompt "你可以帮我吗？谢谢你的帮忙！议地点在哪里？我昨天看了一部电影。"

basically just include a few sentences with punctuation included.

[SteveTanggithub]

thank you! We now get a new issue. We input a wav file about 3 seconds but we got this:

it output some words don't exist in the speech and expand the duration to 33 seconds! We use main.cpp in the main fold. What's the problem? the input audio must longer than 30 seconds? or some config setting we did't change?