FBeta update

[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 #55
Fixes #58

Updates the internals of fbeta metric to use the stats score metric internal (similar to precision/recall). In addition to standardizing the interface even more, it also adds support for:

• TopK calculates
• Average='samples'
• ignore_index argument for ignoring specific classes
• mdmc_average for how to aggregate when working with multi dimensional tensors

When everything is standardized to using the stat score metric it will be easier to implement support for non-normalized input (see issue #60 and #74)

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 🙃

[pep8speaks]

Hello @SkafteNicki! Thanks for updating this PR.

• In the file torchmetrics/functional/classification/f_beta.py:

Line 56:13: W503 line break before binary operator

Comment last updated at 2021-03-24 14:57:42 UTC

[Borda]

could you use the shorten typing for torch.Tensor

[Borda]

@SkafteNicki mind check the failing tests :]

AttributeError: 'FBeta' object has no attribute 'average'

[SkafteNicki]

@Borda should be fixed now :]

[codecov]

Codecov Report

Merging #111 (aa0f001) into master (9a73d5e) will decrease coverage by 0.49%.
The diff coverage is 98.14%.

@@            Coverage Diff             @@
##           master     #111      +/-   ##
==========================================
- Coverage   97.25%   96.75%   -0.50%     
==========================================
  Files         140      140              
  Lines        4366     4376      +10     
==========================================
- Hits         4246     4234      -12     
- Misses        120      142      +22     

Flag	Coverage Δ
Linux	78.83% <68.51%> (-1.06%)	⬇️
Windows	78.83% <68.51%> (-1.06%)	⬇️
cpu	96.75% <98.14%> (-0.50%)	⬇️
gpu	96.75% <ø> (-0.50%)	⬇️
macOS	96.75% <98.14%> (-0.50%)	⬇️
pytest	96.75% <98.14%> (-0.50%)	⬇️
python3.6	96.74% <98.14%> (-0.50%)	⬇️
python3.8	96.75% <98.14%> (-0.50%)	⬇️
python3.9	96.75% <98.14%> (-0.50%)	⬇️
torch1.3.1	?
torch1.4.0	95.33% <98.14%> (-0.50%)	⬇️
torch1.8.0	96.75% <98.14%> (-0.50%)	⬇️

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

Impacted Files	Coverage Δ
...hmetrics/functional/retrieval/average_precision.py	80.00% <50.00%> (ø)
torchmetrics/classification/f_beta.py	100.00% <100.00%> (+3.12%)	⬆️
torchmetrics/functional/classification/f_beta.py	100.00% <100.00%> (ø)
torchmetrics/retrieval/retrieval_metric.py	98.00% <100.00%> (ø)
torchmetrics/utilities/data.py	95.31% <100.00%> (ø)
torchmetrics/utilities/checks.py	91.55% <0.00%> (-7.80%)	⬇️
__w/1/s/torchmetrics/utilities/checks.py	91.55% <0.00%> (-7.80%)	⬇️
...s/torchmetrics/functional/classification/f_beta.py	100.00% <0.00%> (ø)
__w/1/s/torchmetrics/classification/f_beta.py	100.00% <0.00%> (+3.12%)	⬆️

---

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 9a73d5e...aa0f001. Read the comment docs.

[Borda]

cool, the last missing is docs, some formatting issue...