Patricio Cerda-Mardini

This is now yielding an informative error message, closing.

We have an open PR that is stale but very close to completion: #8180. Perhaps we should look into adding the missing tests.

@parthiv11 copying comment from #8861 here, for reference. Sorry about the delay: > we'll kindly ask that you please work on this one in the same branch and pull request...

This confusion is due to a lack of documentation on the underlying ML engine, Lightwood. This is expected (although unintuitive!) behavior. The `0.0001` confidence score you see is produced by...

Ah, thanks for that link Max, very informative 🔍. > users had troubles with their files: they uploaded files with column names with spacebars, but in query used names without...

@pedrofluxa @dusvyat @tomhuds @QuantumPlumber do you guys have any thoughts on this? I still think we should remove this behavior.

@ZoranPandovski FYI ^

Lightwood can run on both CPU and GPU. The issue here has to do with storing models in one node type then trying to run them on another, so the...

@martyna-mindsdb what kind of mindsdb deployment is this on? cloud?

@yuhuishi-convect indeed. There's the `lightwood/helpers/device` set of methods. I think to reproduce, we'd need a lightwood model that was trained and saved in a GPU-aware environment, then trying to load...