SenseVoice
SenseVoice copied to clipboard
Multilingual Voice Understanding Model
使用了大概800小时左右的中文音频按照readme的步骤进行了微调,测试了中文识别能力后想测试一下对原有的其他语种识别能力是否有影响,结果发现对于日语音频会按照中文进行识别,指定推理时的语种参数也没有效果,请问是因为微调时数据仅有中文造成的影响吗,还是说? 有别的朋友遇到这个问题吗?
### question ```python I first followed the README and ran pip install -r requirements.txt to set up the environment. Then, I continued following the README under Usage → Inference and...
During batch inference with timestamps, a tensor access out-of-bounds issue consistently occurs. It has been identified that when obtaining the timestamp, the input to the `ctc_forced_align` function should be a...
Notice: In order to resolve issues more efficiently, please raise issue following the template. (注意:为了更加高效率解决您遇到的问题,请按照模板提问,补充细节) ## ❓ Questions and Help ### Before asking: 1. search the issues. 2. search the...
the original webui.py cannot run without internet connection. Now, if we already run it once before, it could run offline with models in huggingface cache.
Notice: In order to resolve issues more efficiently, please raise issue following the template. (注意:为了更加高效率解决您遇到的问题,请按照模板提问,补充细节) ## ❓ Questions and Help ### Before asking: 1. search the issues. 2. search the...
Notice: In order to resolve issues more efficiently, please raise issue following the template. (注意:为了更加高效率解决您遇到的问题,请按照模板提问,补充细节) ## ❓ Questions and Help ### Before asking: 1. search the issues. 2. search the...
Does CTC computation have a length limit? I noticed that when the audio is too long, the ending gets truncated.