-
-
Notifications
You must be signed in to change notification settings - Fork 178
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Colbert rerank not working, no attribute "run_context" #91
Comments
I think when no run_context is found, it can't find the index directory. When running from_index, I have to submit the full path to dir. not sure if that's intended. |
You should be able to do this without an index
But for the exception above is given on the line:
Indicating a issue with RAGPretrainedModel class it seems, will take a deeper look soon |
My mistake. I misread your initial comment. |
If I understand correctly, this is intended yes! Maybe it's not super intuitive 🤔 @julianmukaj It seems like this is likely hiding another error, as |
That got past the issue, seems to not be compatible in Windows when not using WSL. Related issue: #14 You can resolve that users issue by installing the Windows C++ Build Tools and adding it to PATH correctly. But then you get to here:
Should be able to resolve this by installing pthreads-w32, but having trouble getting colbert to detect it for now. I think fair to add library is currently not supported for Windows outside of WSL. |
Yeah this is fair. I've juste added a disclaimer to the README to this effect -- windows support right now is a nice to have but sadly not something we can target with current resources |
\appdata\local\programs\python\python39\lib\site-packages\ragatouille\models\colbert.py", line 752, in del
AttributeError: 'ColBERT' object has no attribute 'run_context'
Have tried all versions of RAGatouille, this is on windows and on several Py versions 3.8, 3.9, 3.10 have been tested.
I saw issue #40 fixed it for RAGtrainer but this happens when I load RAGPretrainedModel
The text was updated successfully, but these errors were encountered: