Nicki Skafte Detlefsen

Hi @adamklie, Could you please provide us with some testing data that creates this discrepancy between our implementation and sklearns?

After looking at the code again, it seems clear to be that the difference is simply due to the input order. Sklearn expects input as `metric(target, preds)` whereas we in...

Hi @richarddwang, To make the debugging easier, could you provide a fully reproducible script ?

Issue will be fixed by classification refactor: see this issue https://github.com/Lightning-AI/metrics/issues/1001 and this PR https://github.com/Lightning-AI/metrics/pull/1195 for all changes Small recap: This issue describes that `jaccard_index` is wrongly calculated in the...

@ragavvenkatesan do we have some kind of reference implementation of this metric (sklearn, scipy, something else)? That is normally what we need for testing to secure that the implementation is...

Hi @ragavvenkatesan, It seems like we can compare against: https://kornia.readthedocs.io/en/latest/_modules/kornia/losses/total_variation.html

Issue will be fixed by classification refactor: see this issue https://github.com/Lightning-AI/metrics/issues/1001 and this PR https://github.com/Lightning-AI/metrics/pull/1195 for all changes Small recap: This issue describe that metric `stat_scores` is not correctly working...

@stancld sounds great :]

@MeganClark-Robinson: * 'global' is probably the easiest to understand. This just means that all additional dimensions are just considered extra observations. So if you have tensor with provabilities that have...

@AndreiSiliUiPath thanks for bringing up this issue. I am not sure at the moment if there is a bug here or the result should be like that. I have to...