fix: keras import errors in new versions

[jgongd]

Ticket

MD-358

Description

Update the import paths of some modules in tensorflow and Keras.

Test Plan

Create experiments and see if they complete successfully:

1. example/computer_vision/iris_tf_keras/const.yaml: Use determinedai/tensorflow-ngc:748dda4 in the experiment config.
2. example/computer_vision/iris_tf_keras/const.yaml: Use determinedai/tensorflow-ngc-hpc:748dda4 in the experiment config.

Checklist

• Changes have been manually QA'd
• User-facing API changes need the "User-facing API Change" label.
• Release notes should be added as a separate file under docs/release-notes/.
  See Release Note for details.
• Licenses should be included for new code which was copied and/or modified from any external code.

[netlify]

✅ Deploy Preview for determined-ui ready!

Name	Link
🔨 Latest commit	1bf8fb0
🔍 Latest deploy log	https://app.netlify.com/sites/determined-ui/deploys/661838c228f29b00087e6415
😎 Deploy Preview	https://deploy-preview-9141--determined-ui.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify site configuration.

[codecov]

Codecov Report

Attention: Patch coverage is 60.00000% with 4 lines in your changes are missing coverage. Please review.

Project coverage is 42.20%. Comparing base (d0497da) to head (1bf8fb0).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #9141      +/-   ##
==========================================
- Coverage   46.54%   42.20%   -4.34%     
==========================================
  Files         755     1147     +392     
  Lines      104838   143603   +38765     
  Branches     2413     2413              
==========================================
+ Hits        48795    60612   +11817     
- Misses      55834    82782   +26948     
  Partials      209      209              

Flag	Coverage Δ
backend	43.76% <ø> (+9.51%)	⬆️
harness	49.53% <60.00%> (-13.41%)	⬇️
web	36.76% <ø> (ø)

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

Files	Coverage Δ
harness/determined/keras/_load.py	87.50% <80.00%> (-1.14%)	⬇️
harness/determined/keras/_tf_keras_trial.py	81.09% <40.00%> (-0.48%)	⬇️

... and 591 files with indirect coverage changes

[MikhailKardash]

we have a code style guideline where we don't want to use object names directly without modules.

Can you change both of these imports to something like:
from tensorflow.python.trackable import autotrackable as tf_tracking
from tensorflow.python.training.tracking import tracking as tf_tracking

Then you can reference the object as tf_tracking.AutoTrackable

[jgongd]

Thanks. I hadn't consider renaming the module, so I can now use new_module_name.object as the type and don't need noqa: I2041 in importing the current version tensorflow.

[MikhailKardash]

For the test plan, I think it's sufficient to test only the TensorFlow image changes based on what is changed in this PR.

I would update the test plan to test the NGC image and the 2.12 TF image.

[MikhailKardash]

LGTM