chore: Add Speech Recognition with DDP Example

[zren11]

What this PR does / why we need it:
This is an simple Speech Recognition example for Issue 2040.

Key Features:

• Complete end-to-end workflow in example.ipynb - from data preparation to distributed training
• PyTorch DDP integration with Kubeflow Trainer for multi-node speech recognition
• Transformer-based model (4 layers, 4 attention heads) for audio classification
• Use Google Speech Command dataset - 35 words, 105K+ 1-second audio files for classification
• Local to distributed progression - starts with local training, scales to Kubernetes

Details can be found in PR's README.md file.

Which issue(s) this PR fixes (optional, in Fixes #<issue number>, #<issue number>, ... format, will close the issue(s) when PR gets merged):
Fixes #Issue 2040 Issue 2040.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[google-oss-prow]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please assign andreyvelich for approval. For more information see the Kubernetes Code Review Process.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[coveralls]

Pull Request Test Coverage Report for Build 17773026722

Warning: This coverage report may be inaccurate.

This pull request's base commit is no longer the HEAD commit of its target branch. This means it includes changes from outside the original pull request, including, potentially, unrelated coverage changes.

• For more information on this, see Tracking coverage changes with pull request builds.
• To avoid this issue with future PRs, see these Recommended CI Configurations.
• For a quick fix, rebase this PR at GitHub. Your next report should be accurate.

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 55.137%

---

Totals
Change from base Build 17772927799:	0.0%
Covered Lines:	1084
Relevant Lines:	1966

---

💛 - Coveralls

[zren11]

@andreyvelich I notice the GPU E2E test failed. Is there anything I can do to fix it? Thanks!

[jaiakash]

/ok-to-test

[andreyvelich]

@andreyvelich I notice the GPU E2E test failed. Is there anything I can do to fix it? Thanks!

Can you rebase your PR again please ?

[andreyvelich]

You can remove all of your files and keep only Jupyter Notebooks.
You can check out this example for knowledge distillation BERT fine-tuning: https://github.com/kubeflow/trainer/blob/master/examples/pytorch/question-answering/fine-tune-distilbert.ipynb
In your Notebook you can reference to the getting started guide which explains how to deploy kind cluster and torch-distributed Runtime.

[jaiakash]

/retest

[jaiakash]

Hi @zren11 I see still the few tests are failing. Can you please rebase the PR!

[jaiakash]

Hi @zren11 are you still working on this, if yes can you please rebase the PR.

[Snehadas2005]

Hi @jaiakash and @andreyvelich, I have pushed the updates to finalise this PR under PR #3063
The example is now aligned with the current Trainer example patterns, uses a notebook based format, and addresses the earlier feedback.
Please take a look when convenient. Happy to make any further adjustments if needed.

[ananyajain10]

Hi @jaiakash , @andreyvelich ,

I’ve raised a PR with the required changes for this task:
#3156

Would appreciate your review. Thanks!