refactor: formalize artifact saving

[gsprochette]

Description

Formalize artifact saving and expose it to more algorithms, and add artifact hooks into Smash config

• Generalized artifact saving/loading: Extracted and formalized the artifact-saving logic so it is no longer tied specifically to the Torch Compile workflow. Other Pruna algorithms can now register and reuse the same standardized artifact save/load utilities.
• Config-driven hooks: Extended the Smash configuration to include save_artifacts_fns and load_artifacts_fns, working in a way similar to save_fns and load_fns
• Improved composability: This structure makes it easy to save artifacts (such as parameters estimated from a warmup, or checkpoints for finetuning or distillation).

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

How Has This Been Tested?

Checklist

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

Additional Notes

[cursor]

Comment @cursor review or bugbot run to trigger another review on this PR

[cursor]

Loading old models with torch artifacts crashes

High Severity

Models saved before this refactoring have torch_artifacts as an entry in load_fns. The old loading code would detect and remove this entry before processing. With the new code, LOAD_FUNCTIONS[smash_config.load_fns[0]] raises a KeyError when torch_artifacts is the first entry, since torch_artifacts was removed from LOAD_FUNCTIONS without adding migration handling in load_from_json.