fix(layers): Update Conv2D docstring to clarify numerical precision across backends

[harshaljanjani]

• Make it clear that, contrary to what the Keras 3 docs state, Conv2D operations may exceed the documented 1e-7 precision difference across backends.
• Document that large convolutions can show notable variations due to accumulated floating-point operations.

Fixes #20804.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 82.24%. Comparing base (c04cf9d) to head (3953bf4).
Report is 2 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master   #20867   +/-   ##
=======================================
  Coverage   82.24%   82.24%           
=======================================
  Files         561      561           
  Lines       52633    52633           
  Branches     8137     8137           
=======================================
  Hits        43288    43288           
  Misses       7340     7340           
  Partials     2005     2005           

Flag	Coverage Δ
keras	82.05% <ø> (ø)
keras-jax	64.22% <ø> (ø)
keras-numpy	59.03% <ø> (-0.01%)	⬇️
keras-openvino	32.52% <ø> (ø)
keras-tensorflow	64.85% <ø> (ø)
keras-torch	64.27% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

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

[fchollet]

Thank you!