CODEWITHSUNDEEP
pythonindexingvespavector-database
DVS
DVS

Reputation: 9

the documented example of vespa is not working

I'm trying to replicate one of the documented examples. But it is not working at all. Kind of weird as the documented example is throwing off these Error.

the issue is at max_sim RuntimeError: Deployment failed, code: 400, message: {'error-code': 'INVALID_APPLICATION_PACKAGE', 'message': "Invalid application: In schema 'testapp', rank profile 'combined_ranking': The function 'context_max_sim' is invalid: max_sim is invalid: Can not join tensor(qt{},x[128]) and tensor(context{},dt{},x[16]): Unequal dimension x in tensor(qt{},x[128]) and tensor(context{},dt{},x[16])"}


            )
        )```

Upvotes: 0

Views: 143

Answers (1)

thomasht86
thomasht86

Reputation: 51

I assume you are referring to this example: https://pyvespa.readthedocs.io/en/latest/examples/colbert_standalone_long_context_Vespa-cloud.html.

I can confirm that this runs, but it seems you are modifying quite a bit. Also, in your error message, you refer to Rank Profile combined_ranking, while in your snippet, you have hybrid_ranking. Are you sure they are the same?

I can strongly recommend using the Tensor Playground for debugging tensor expressions.

Upvotes: 3

Related Questions