Fix the dependencies of LPIPS metric

[tanguymagne]

What does this PR do?

This PR doesn't solve any issue, as it is a very small fix.

Since #1575, the LPIPS metric does not rely anymore on the third-party package PerceptualSimilarity but now relies on torchvision. The changes were not reflected entirely in the doc and in the tests.

Before submitting

• Was this discussed/agreed 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? --> Not necessary.

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 🙃

---

📚 Documentation preview 📚: https://torchmetrics--2230.org.readthedocs.build/en/2230/

[Borda]

this does not seems to be right, check:

torchmetrics/tests/unittests/image/test_lpips.py

Line 19 in bcf7867

from lpips import LPIPS as LPIPS_reference # noqa: N811

[tanguymagne]

I have made the requested changes. I hope I understood correctly what was requested.
I am wondering whether or not it is a problem to have a dependency on the unmaintained third-party package lpips for testing ?

[codecov]

Codecov Report

Merging #2230 (dea4c8e) into master (c35a2fb) will decrease coverage by 47%.
Report is 1 commits behind head on master.
The diff coverage is 67%.

Additional details and impacted files

@@           Coverage Diff            @@
##           master   #2230     +/-   ##
========================================
- Coverage      87%     40%    -47%     
========================================
  Files         293     293             
  Lines       16423   16425      +2     
========================================
- Hits        14278    6641   -7637     
- Misses       2145    9784   +7639