Kevin Hu
Kevin Hu
> > Could you mock up a sample that will fail. > > @KevinHuSh This is a sample file. > > [1.xlsx](https://github.com/user-attachments/files/16050142/1.xlsx) This file can be parse smoothly on demo...
> 请问一下，你这个问题解决了吗？我也遇到了这个问题。 Could you elabrate on your issue? Attach the sample file if it's okay for you.
Method 'Table' is for data dumped from DB, so, you'd better parse it by 'General'.
Yes, it doesn't make sense.
In version 0.7.0, the default embedding changed a bit, I'm afraid you need to redo parsing your documents using a valid embedding model.
Go to system setting > Model Providers > set a valid key of your selected models.
@ericwu0930 The median of retrieval time for a query is 6009 ms isn't it? Did you use default embedding model which was deployed on CPU or other embedding SAAS?
@ericwu0930 Did you utilize rerank model?
You can see the stats since 'yesterday'.
What about disable this? 