feat(semseg): allow model customization

[borisdayma]

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?
  not applicable
• Did you write any new necessary tests?
  not applicable
• If you made a notable change (that affects users), did you update the CHANGELOG?

What does this PR do?

Allow more customization to the semantic segmentation example:

• test set has been replace with a validation set to measure model performance
• both training and validation losses are logged
• unet allows for more customization: number of layers and number of features
• masks have been updated to be compatible with the most recent version of the dataset which contains more labels (additional labels are just ignored)
• script allows for other arguments: distributed backend, 16 bit, number of layers and features for unet as well as type of upsampling, gradient accumulation, number of epochs, logging on W&B

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 @borisdayma! Thanks for updating this PR.

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

Comment last updated at 2020-04-02 05:56:13 UTC

[borisdayma]

Note: this example requires #1311

[codecov]

Codecov Report

Merging #1338 into master will not change coverage.
The diff coverage is n/a.

@@          Coverage Diff           @@
##           master   #1338   +/-   ##
======================================
  Coverage      92%     92%           
======================================
  Files          62      62           
  Lines        3239    3239           
======================================
  Hits         2966    2966           
  Misses        273     273

[borisdayma]

The build error related to "examples" mentions problem of relative import which was already there before. Not sure if you want to change it. I think it's clean this way.

[mergify]

This pull request is now in conflict... :(

[williamFalcon]

@borisdayma #1311 has been merged! no longer a blocker. please rebase and resubmit PR.

[borisdayma]

I think this PR can now be closed in favor of #1371.