Ingest metrics from autoevaluator #2

by lewtun - opened

For some reason, the leaderboard doesn't display the results from the evaluation service - example here.

Perhaps the issue is that we now report the same metrics (accuracy & F1 score) twice (verified vs self-reported)?

I'm on it!

Thanks

Oh yes probably because of what you mention.

I've pushed a fix. Let me know if there are any other issues.

Closing because I think we're good now

Tristan changed discussion status to closed