Proper multilabel support for confmat

[SkafteNicki]

Before submitting

• Was this discussed/approved via a Github issue? (no need for typos and docs improvements)
• Did you read the contributor guideline, Pull Request section?
• Did you make sure to update the docs?
• Did you write any new necessary tests?

What does this PR do?

Fixes #100
Currently using confusion matrix with multilabel data will smash it all together as one big binary classification problem, which is really not wanted. Instead we should probably go for something similar to what sklearn does in that case (https://scikit-learn.org/stable/modules/generated/sklearn.metrics.multilabel_confusion_matrix.html). Instead of having a specialized metric for this case, this PR adds a is_multilabel flag and implements the same logic as sklearn.

PR review

Anyone in the community is free to review the PR once the tests have passed.
If we didn't discuss your PR in Github issues there's a high chance it will not be merged.

Did you have fun?

Make sure you had fun coding 🙃

[codecov]

Codecov Report

Merging #134 (484b9fb) into master (19b77cc) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master     #134   +/-   ##
=======================================
  Coverage   96.76%   96.77%           
=======================================
  Files         156      156           
  Lines        4826     4837   +11     
=======================================
+ Hits         4670     4681   +11     
  Misses        156      156           

Flag	Coverage Δ
Linux	79.26% <93.33%> (-0.02%)	⬇️
Windows	79.26% <93.33%> (-0.02%)	⬇️
cpu	96.77% <100.00%> (+<0.01%)	⬆️
gpu	96.77% <ø> (+0.01%)	⬆️
macOS	96.77% <100.00%> (+<0.01%)	⬆️
pytest	96.77% <100.00%> (+<0.01%)	⬆️
python3.6	96.77% <100.00%> (+0.01%)	⬆️
python3.8	?
python3.9	?
torch1.3.1	95.57% <100.00%> (+0.01%)	⬆️
torch1.4.0	?
torch1.8.1	96.77% <100.00%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
torchmetrics/classification/confusion_matrix.py	100.00% <100.00%> (ø)
...rics/functional/classification/confusion_matrix.py	100.00% <100.00%> (ø)
torchmetrics/metric.py	95.00% <0.00%> (-0.08%)	⬇️
torchmetrics/utilities/prints.py	100.00% <0.00%> (ø)
.../s/torchmetrics/classification/confusion_matrix.py	100.00% <0.00%> (ø)
...rics/functional/classification/confusion_matrix.py	100.00% <0.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 19b77cc...484b9fb. Read the comment docs.

[Borda]

just not sure about the is_multilabel sound unnatural to me... (as an argument, it would be fine as property or attribute)

[SkafteNicki]

@Borda I have changed the arg to just multilabel, as you are right it makes more sense