User bug: missing secrets for private modellab files when running public benchmark

[bkorycki]

Sachi reports an error when running the command poetry run modelbench benchmark -m 10 --plugin-dir plugins/localModel/ --sut local_sut

I was trying to run modelbench, I think there were some latest changes in the last couple of days. When I try to run a sample of the benchmark, I get this:

raise MissingSecretValues.combine(missing_secrets)
modelgauge.secret_values.MissingSecretValues: Missing the following secrets:
scope='modellab_files' key='token' instructions='Please ask MLCommons admin for permission.'

I hadn't seen this before, is this for the latest practice set?
i added a new plugin - to read a local model. and it works with the "poetry run modelgauge run-csv-items sample.csv --sut local_sut

[mayank-soni]

+1 for facing this same issue with just a default poetry run modelbench benchmark -m 10 as per the testing instructions on https://github.com/mlcommons/modelbench/blob/main/README.md

[bkorycki]

Hi @mayank-soni, this bug was actually resolved already. Can you confirm that you have the latest version of modelbench checked out?