Hi,
when I use Kimi infer framework to do local test, I met a problem when audio (.wave) length over 30s as input of kimi-audio-7b, No response and exception. even no error message from console. It has nothing to do with GPU and CPU as it still has enough resources there. I'm not sure if I missed some configurations or I have to split audio to chunks (for example: chunk _length_s = 30) before entering Kimi LLM. please help me check it, thank you.
Output in console:
音频验证通过：39.36秒，24000Hz，单声道
Generating tokens: 1%|▎ | 42/6827 [00:02<06:50, 16.54it/s]
--- no output from here
Regards,
Xing
Hi,
when I use Kimi infer framework to do local test, I met a problem when audio (.wave) length over 30s as input of kimi-audio-7b, No response and exception. even no error message from console. It has nothing to do with GPU and CPU as it still has enough resources there. I'm not sure if I missed some configurations or I have to split audio to chunks (for example: chunk _length_s = 30) before entering Kimi LLM. please help me check it, thank you.
Output in console:
音频验证通过：39.36秒，24000Hz，单声道
Generating tokens: 1%|▎ | 42/6827 [00:02<06:50, 16.54it/s]
--- no output from here
Regards,
Xing