Closed sjawhar closed 2 days ago
@maxisawesome can you please take a look?
CC: @dakinggg
YES!! It's finally working! The logging, that is. Not my fine-tuning, that's still garbage. But I can finally start debugging!
Looks great other than the one comment I left! Thanks for fixing this for non-generative evals.
@sjawhar if you can fix lint + unit tests (the CPU ones) that would be awesome! then we can merge
Anything else I can do to help get this merged?
@sjawhar sorry about that, will take a look this week!
LGTM, please update the PR title and description to reflect the changes in this PR. Thank you!
Done
@sjawhar looks like the tests failed with the recent change
Adds the ability to log text output using
EvalOutputLogging
in non-metrics/ICL use cases. This appears as a newoutputs
key in the logged dictionary, which is set tostats.outpus
and de-tokenized when the model is aHuggingFaceModel
and the dataset has a tokenizer.