Mps backened support

[jrzaurin]

The name of this PR is a bit misleading, since a number of things have happened here

1. Added support for MPS backened
2. Added a series of models to the rec module: DCN, DCNv2, GDCN, AutoInt, AutoIntPlus
3. Added a DIN preprocessor
4. Reviewed the docs
5. Reviewed the examples
6. Other (minor and not so minor) fixes

[codecov]

Codecov Report

Attention: Patch coverage is 95.03476% with 50 lines in your changes missing coverage. Please review.

Project coverage is 94.87%. Comparing base (9df6585) to head (278c72f).
Report is 25 commits behind head on master.

Files with missing lines	Patch %	Lines
pytorch_widedeep/utils/general_utils.py	38.70%	19 Missing ⚠️
pytorch_widedeep/dataloaders.py	80.55%	7 Missing ⚠️
pytorch_widedeep/preprocessing/din_preprocessor.py	97.26%	5 Missing ⚠️
pytorch_widedeep/training/trainer.py	93.33%	5 Missing ⚠️
pytorch_widedeep/metrics.py	98.33%	3 Missing ⚠️
pytorch_widedeep/losses.py	81.81%	2 Missing ⚠️
pytorch_widedeep/models/model_fusion.py	96.42%	2 Missing ⚠️
pytorch_widedeep/models/rec/autoint_plus.py	97.26%	2 Missing ⚠️
pytorch_widedeep/tab2vec.py	66.66%	2 Missing ⚠️
pytorch_widedeep/training/bayesian_trainer.py	77.77%	2 Missing ⚠️
... and 1 more

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #235      +/-   ##
==========================================
+ Coverage   94.80%   94.87%   +0.07%     
==========================================
  Files         121      126       +5     
  Lines        7579     8277     +698     
==========================================
+ Hits         7185     7853     +668     
- Misses        394      424      +30     

Flag	Coverage Δ
unittests	94.87% <95.03%> (+0.07%)	⬆️

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

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.