Kuang Ru

Please see codes in wsj local/pytorchnn. It's another nnlm rescore backend.

I'm sorry that I am not in this direction now. The conclusion of the previous attempt is that it will obviously reduce server throughput. So keeping the existing status is...

> 先配置logging再import lac，logging配置可生效，具体原因还不太清楚，可能与logging实现有关 > > ```python > import logging > logging.basicConfig(filename='testlog.log', filemode="a", format="%(asctime)s %(name)s:%(levelname)s:%(message)s", datefmt="%Y-%m-%d %H:%M:%S",level=logging.DEBUG) > > from LAC import LAC > > logging.info('test lac log') > ``` 应该是在某个地方初始化了root...

Resolved. The wrong output shape is provided. It may be caused by convert.py

Sure, here are these two files https://gist.github.com/yzmyyff/36fff2a3be9b57870034b0787e70f5ee How about raising a warning if no shape is provided.

I think I don't get your point. Tritonserver supports Auto-Generated Model Configuration. So I just remove the output in the template. Do you mean that we must define output shape...

Why not use HTTP API? ```bash curl localhost:8000/v2/models//config ``` Then we can completely avoid processing the model details.

Great! This PR can be closed If you're working on that.

I can't find alignment_log_likelihood.csv in MFA 3. does anyone know how it was generated?

我也遇到了一样的问题, 如果把funasr版本降级到1.0.17就没有异常了.