PyTorch model extractor

[mastoffel]

• adds a function to extract PyTorch models from AutoEmulate emulators

• emulators can be pipelines, MultiOutputRegressors etc., this function checks all the options and extracts the underlying PyTorch model where possible and throws an error for other models

• it also gives a message saying that datapreprocessing is better turned off when doing this and has to be done manually (as it can't be attached to the PyTorch model like it can be to a sci-kit learn model using a pipeline

• it returns the model in eval mode

• does not yet include other objects as discussed in Add "fit only pytorch models" flag #291 . Maybe we leave that to the next PR?

[github-actions]

Coverage report

Click to see where and how coverage changed

File	Statements	Missing	Coverage	Coverage (new stmts)	Lines missing
autoemulate
utils.py					426, 435, 449
autoemulate/emulators
conditional_neural_process.py
gaussian_process.py
gaussian_process_mt.py
tests
test_pytorch_utils.py
Project Total

This report was generated by python-coverage-comment-action

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 96.47059% with 3 lines in your changes missing coverage. Please review.

Project coverage is 94.22%. Comparing base (7a4dc72) to head (40dbed2).
Report is 5 commits behind head on main.

Files with missing lines	Patch %	Lines
autoemulate/utils.py	85.71%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #298      +/-   ##
==========================================
+ Coverage   94.20%   94.22%   +0.02%     
==========================================
  Files          62       63       +1     
  Lines        3606     3691      +85     
==========================================
+ Hits         3397     3478      +81     
- Misses        209      213       +4     

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

[marjanfamili]

Thanks a lot for this PR Martin. It looks great. Would this functionality be necessary if we change everything to PyTorch? is this a temporary measure until we do ?

[mastoffel]

Thanks a lot for this PR Martin. It looks great. Would this functionality be necessary if we change everything to PyTorch? is this a temporary measure until we do ?

@marjanfamili, I think it wouldn't be necessary if everything is changed to PyTorch! So, yes, temporary.

[marjanfamili]

Looks great, thanks