Add missing default in docstring

Fix formatting

Switch order of torch depending on operating system

Fix issue with torch imported before Julia init

Add determinism-based scikit-learn tests back

Add deterministic mode with new SR version

Fixed weight checking

Supress progress output for tests

Fix to pass test_custom_operator

Ensure iteration completes for multioutput test

Added validation for weights

Passed check_dict_unchanged

More helpful error message for sklearn tests

Soft fail tests which require determinism

Don't check array if set to None

Updated parameter validation

Add warm_start

Cleaned test and docstring

New updated checkpoint loading and tests

Added pickle support

Added control of random_state for numpy and julia

Merge branch 'master' into refactor-PySRRegressor

TODO for refactoring Julia interface

Check shape length after validation

Automatically skip pickling tests

Move binary_operator assertion to after they are defined

Make __init__ not modify parameters again

Fix parameter name in docs

Fix occurrence of size warning

Ensure large size test quits early even with fail

Print details of scikit-learn checks and failures

Fix ordering of torch import

Order torch imports after Julia init

Fix variable selection for PyTorch

Add feature selection for PyTorch

Fix feature selection for JAX export

Add test for feature selection in JAX output

More descriptive name for warm_start test

Move shape checking to fit()

Remove additional changes to internal parameters

Remove misplaced parameter changes

Create function to setup equation file during fit

Process params in __init__ instead of fit

Move shape checking to _validate_fit_params

Change back default test kwargs

Speed up scikit-learn tests

Move scikit-learn test to miscellaneous tests

Hotfix for selection test

Add ScikitLearn check_estimator test

Add another selection test, but with numpy arrays