Fix model offloading and training tests + prevent examples timeout

[GiGiKoneti]

What does this PR do?

This PR fixes three pre-existing bugs in model offloading, training tests, and example test runs that were causing CI failures:

1. AutoencoderVidTok return format mismatch:

   • Modified AutoencoderVidTok.forward to return (dec,) when return_dict=False, aligning it with the standard VAE return contract in Diffusers.
   • Unskipped test_outputs_equivalence since the outputs match correctly now.

2. AutoencoderDC mixed precision skip:

   • Added a try-except block to catch RuntimeError: "GET was unable to find an engine to execute this computation" and call pytest.skip when cuDNN cannot find matching computation engine configs.

3. Examples timeout and distributed launch mitigation:

   • Refactored run_command in examples/test_examples_utils.py to use subprocess.run with a configurable timeout parameter (default 300s) to prevent tests from hanging indefinitely.
   • Appended --num_processes 1 --num_machines 1 to ExamplesTestsAccelerate launch arguments to prevent distributed launch deadlocks on single-device/CPU CI runners.

Fixes #14090

Before submitting

• Did you use an AI agent (Claude Code, Codex, Cursor, etc.) to help with this PR? If so:
  • Did you read the Coding with AI agents guide?
  • Did you self-review the diff against .ai/review-rules.md?
• Did you read the contributor guideline?
• Did you read our philosophy doc? (important for complex PRs)
• Was this discussed/approved via a GitHub issue or the forum? Please add a link to it if that's the case.
  • Approved via issue: Fix model offloading and training tests + prevent examples timeout #14090
• Did you make sure to update the documentation with your changes?
• Did you write any new necessary tests?
• Are you the author (or part of the team) of the model/pipeline?

Who can review?

@sayakpaul @DN6 @pcuenca