Pretrained flag and resnet50 pretrained weights

[ananyahjha93]

What does this PR do?

1. Change pretrained flag in ImageClassifier to load self-supervised pretrained weights for backbones
2. Add pretrained weights of simclr, swav, barlow-twins for resnet50 backbone from VISSL.
3. Removes dependency on bolts for pretrained SSL weights.

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 your PR does only one thing, instead of bundling different changes together?
• Did you make sure to update the documentation with your changes?
• Did you write any new necessary tests? [not needed for typos/docs]
• Did you verify new and existing tests pass locally with your changes?
• If you made a notable change (that affects users), did you update the CHANGELOG?

PR review

• Is this pull request ready for review? (if not, please submit in draft mode)

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 @ananyahjha93! Thanks for updating this PR.

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2021-07-12 14:39:26 UTC

[codecov]

Codecov Report

Merging #560 (d15d66b) into master (48bdfd8) will decrease coverage by 0.06%.
The diff coverage is 75.86%.

@@            Coverage Diff             @@
##           master     #560      +/-   ##
==========================================
- Coverage   91.33%   91.27%   -0.07%     
==========================================
  Files         113      113              
  Lines        7202     7206       +4     
==========================================
- Hits         6578     6577       -1     
- Misses        624      629       +5     

Flag	Coverage Δ
unittests	91.27% <75.86%> (-0.07%)	⬇️

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

Impacted Files	Coverage Δ
flash/image/classification/model.py	78.84% <28.57%> (-7.83%)	⬇️
flash/image/backbones.py	86.00% <90.90%> (-0.41%)	⬇️

---

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 48bdfd8...d15d66b. Read the comment docs.

[ethanwharris]

Looking good 😃 I think there's some slightly faulty logic in the load function, maybe some tests for different pretrained configurations would be good

[ananyahjha93]

@ethanwharris is there a way to cache the pretrained weights? Adding tests for pretrained configurations will add significant time to tests since large models need to be downloaded before testing. Hence I wasn't sure to include this.

[ethanwharris]

@ananyahjha93 I think they will get cached automatically but not sure. We should still unit test the function though, you could use patch to prevent things from being downloaded if it is an issue

[ethanwharris]

Awesome, LGTM 😃