Knowledge base vectorization failure issue, I need an answer, is this a bug or something else?

OMate
1

I just feel very strange. Enabling memory saving to the knowledge base is normal, but if I want to modify the knowledge base and remove a certain memory, doesn’t it prompt for re-vectorization? Most of the time it fails. Even if I don’t change anything and re-vectorize, it keeps failing. I have tried the steps in the video tutorial, and there is no problem with that.

Vectorization failed: type ‘int’ is not a subtype of type ‘double’ in type cast

2

The model I am using is Silicon-based Fluid’s Qwen/Qwen3-Embedding-8B

3

Damn, I often tried switching the model to BAAI/bge-m3 as in the video and found the problem was solved, so is it incompatible with Qwen/Qwen3-Embedding-8B?

4

It seems to only support this one vector model, which doesn’t feel very useful.

5

This topic was automatically closed 3 hours after the last reply. New replies are no longer allowed.