[Feature] Load partially instantiated state-dict

[vmoens]

Summary:

Allows to load state_dicts from disk when the saved copy contains more elements that the pre-populated copy.

Test plan:
TODO

Fixes #{issue number}
Closes #101

[codecov]

Codecov Report

Merging #103 (5501d83) into main (a11364c) will decrease coverage by 6.18%.
The diff coverage is 80.85%.

@@            Coverage Diff             @@
##             main     #103      +/-   ##
==========================================
- Coverage   90.40%   84.22%   -6.19%     
==========================================
  Files          25       25              
  Lines        2346     2384      +38     
==========================================
- Hits         2121     2008     -113     
- Misses        225      376     +151     

Impacted Files	Coverage Δ
torchsnapshot/flatten.py	86.73% <66.66%> (-3.63%)	⬇️
torchsnapshot/io_preparer.py	92.55% <72.72%> (-1.07%)	⬇️
torchsnapshot/snapshot.py	92.75% <95.23%> (-3.03%)	⬇️
torchsnapshot/storage_plugins/gcs.py	0.00% <0.00%> (-78.90%)	⬇️
torchsnapshot/storage_plugins/s3.py	25.00% <0.00%> (-65.00%)	⬇️
torchsnapshot/storage_plugin.py	43.75% <0.00%> (-21.88%)	⬇️
torchsnapshot/memoryview_stream.py	63.33% <0.00%> (-6.67%)	⬇️
torchsnapshot/io_types.py	79.68% <0.00%> (-3.13%)	⬇️
torchsnapshot/scheduler.py	91.11% <0.00%> (-2.23%)	⬇️
... and 2 more

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[yifuwang]

Hi @vmoens, thank you so much for raising the issue and drafting a proposal!

Previously, in order to reduce the memory footprint on load, we would find the tensors in the target module's state dict, in-place restore them, and use them to build a state dict for .load_state_dict().

Your proposal made us realize that this is not the right approach. Instead, we should always rebuild the original state dict and reuse the tensor in the target's module's state dict in addition. This way we can achieve the same memory saving while making the behavior much more general.

[facebook-github-bot]

Hi @vmoens!

Thank you for your pull request.

We require contributors to sign our Contributor License Agreement, and yours needs attention.

You currently have a record in our system, but the CLA is no longer valid, and will need to be resubmitted.

Process

In order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA.

Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with CLA signed. The tagging process may take up to 1 hour after signing. Please give it that time before contacting us about it.

If you have received this in error or have any questions, please contact us at [email protected]. Thanks!