Leandro von Werra

After a few hours of my life dedicated to finding the reason the tests suddenly fail I figured out the issue: Transformers made a release so the new example scripts...

So this would be quite a deviation from the current status of metrics: most of them are wrappers around existing libraies. Just to name a few: - "accuracy", "recall", "f1"...

```python TypeError: _compute() missing 1 required positional argument: 'model_id' ``` The issue is that the widget only creates inputs for the features. I think we should add a default model...

Yep, I took that as inspiration :) I'll flesh out the PR for review.

Thanks @Rocketknight1! That looks pretty much like what we want. Maybe we can/need change two things: 1. Pass an evaluate metric instead of a callable and add use the `prediction_processor`...

Hi @PierreColombo These metrics look interesting and I think they would be great community metrics. You could add them with the CLI: ```bash evaluate-cli create "Bary Score" ``` This will...

Hi @PierreColombo, I added a skeleton for Bary Score here: https://huggingface.co/spaces/lvwerra/bary_score/tree/main From the template only the `bary_score.py` needs to be adapted and I copied your class to `score.py`. What's missing...

Hi @LecbychMichal could you check whether this issue was resolved with the latest version? If not could you provide the version you are using?

I feel like the former is probably best addressed in the preprocessing: It's literally a onliner to merge to text column into one (`ds = ds.map(lambda x: {"text": x[col1] +...

Looks great! Thanks for adding this, I left a a few minor comments. I agree with @ola13 that splitting such a PR in two would make it much easier to...