Clarification: new mean option is synonym with sum_over_batch_size in loss function base class

[jm-willy]

Since sum_over_batch_size actually is mean, mean option was added as equivalent of sum_over_batch_size to explicitly and clearly state what sum_over_batch_size does, under the expected name mean, while keeping sum_over_batch_size for backwards compatibility. See also #18818, where the problem with the current name is noted.

Also, why is done like this:

loss = ops.sum(values)
loss /= ops.cast(
    ops.prod(ops.convert_to_tensor(ops.shape(values), dtype="int32")),
    loss.dtype,
)

if there's the clearer and shorter keras mean?
loss = ops.mean(values, axis=None, keepdims=False)

Please take it, sum_over_batch_size is confusing, I already implemented mean when I realized it was already done under a different name.

Edit: already undid formating while keeping the typo fixes see last commit: 19933d9

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 78.87%. Comparing base (713382b) to head (f81287f).
Report is 3 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master   #20352   +/-   ##
=======================================
  Coverage   78.87%   78.87%           
=======================================
  Files         512      512           
  Lines       49264    49264           
  Branches     7953     7953           
=======================================
  Hits        38859    38859           
  Misses       8546     8546           
  Partials     1859     1859           

Flag	Coverage Δ
keras	78.74% <ø> (ø)
keras-jax	62.40% <ø> (ø)
keras-numpy	57.42% <ø> (ø)
keras-tensorflow	63.64% <ø> (ø)
keras-torch	62.37% <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]

if reduction in ("mean", "sum_over_batch_size"):

[fchollet]

Same here