[feat] Add backbone API

[tchaton]

What does this PR do?

@ImageClassifier.backbones(name="username/resnet18")
def fn_resnet(pretrained: bool = True):
    model = torchvision.models.resnet18(pretrained)
    backbone = nn.Sequential(*list(model.children())[:-2])
    num_features = model.fc.in_features
    # backbones need to return the num_features to build the head
    return backbone, num_features

print(ImageClassifier.available_backbones())

model = ImageClassifier(backbone="username/resnet18", num_classes=datamodule.num_classes)

Fixes #167

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)

[codecov]

Codecov Report

Merging #204 (41228ba) into master (6a4948a) will increase coverage by 0.21%.
The diff coverage is 92.12%.

@@            Coverage Diff             @@
##           master     #204      +/-   ##
==========================================
+ Coverage   86.31%   86.53%   +0.21%     
==========================================
  Files          56       57       +1     
  Lines        2674     2732      +58     
==========================================
+ Hits         2308     2364      +56     
- Misses        366      368       +2     

Flag	Coverage Δ
unittests	86.53% <92.12%> (+0.21%)	⬆️

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

Impacted Files	Coverage Δ
flash/data/utils.py	96.84% <ø> (-0.04%)	⬇️
flash/text/classification/model.py	93.75% <ø> (-0.19%)	⬇️
flash/vision/backbones.py	81.15% <77.77%> (-3.78%)	⬇️
flash/data/base_viz.py	98.24% <92.85%> (+2.65%)	⬆️
flash/core/model.py	93.97% <100.00%> (+0.26%)	⬆️
flash/core/registry.py	100.00% <100.00%> (ø)
flash/data/data_module.py	78.60% <100.00%> (+0.11%)	⬆️
flash/data/process.py	84.41% <100.00%> (-0.46%)	⬇️
flash/utils/imports.py	100.00% <100.00%> (ø)
flash/vision/__init__.py	100.00% <100.00%> (ø)
... and 7 more

---

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 6a4948a...41228ba. Read the comment docs.

[edgarriba]

minor typing and left some general comments

[carmocca]

Does registering a function to multiple registries or with multiple names work properly? We should test it

[pep8speaks]

Hello @tchaton! Thanks for updating this PR.

• In the file flash/core/registry.py:

Line 120:17: W503 line break before binary operator

Comment last updated at 2021-04-07 23:11:56 UTC

[edgarriba]

🚀🚀