[optimizer] Replace value.nbytes with value.size

[titaiwangms]

To unify the size limitation in terms of the scale.

Passes use tensor size:
https://github.com/onnx/ir-py/blob/a833ab1e178c70046a414b96c1aafbf78a9b4e17/src/onnx_ir/passes/common/constant_manipulation.py#L124

while optimizer uses nbytes, which could potentially confuse users.

[justinchuby]

Thanks!

[justinchuby]

Also need to update https://github.com/titaiwangms/onnx-script/blob/42fafa475d1d4e44b53b3cecc1ab914b07d2c00e/onnxscript/optimizer/_constant_folding.py#L1193-L1198?

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 70.38%. Comparing base (038cac7) to head (39af86f).
Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #2399   +/-   ##
=======================================
  Coverage   70.38%   70.38%           
=======================================
  Files         199      199           
  Lines       25223    25223           
  Branches     2686     2686           
=======================================
  Hits        17754    17754           
  Misses       6540     6540           
  Partials      929      929           

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

[titaiwangms]

I should probably set a compatible default as the previous one (now in different scale...)

[justinchuby]

@copilot can you help update the documentation? (ok maybe not)

[justinchuby]

there is a merge conflict