Added support for Stable Diffusion LCM model

[suryasidd]

Summary

Added initial support for running stable diffusion lcm models.

Test plan

To test this PR, after installing the requirements.
Use export_lcm.py to export
python export_lcm.py --output_dir ./lcm_models --device CPU
Use openvino_lcm.py to infer
python openvino_lcm.py --models_dir ./lcm_models --prompt "Sunset over mountains" --step 4

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/15075

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 2 New Failures

As of commit 43b7bd1 with merge base 4ba2a66 ():

NEW FAILURES - The following jobs have failed:

• pull / test-samsung-models-linux / linux-job (gh)
  RuntimeError: Command docker exec -t 14bb38ff03f46ebd2f773bc837024bbc9d23e337f367271f2c1ad4fa1069f389 /exec failed with exit code 1
• Test CUDA Builds / export-voxtral-cuda-artifact / linux-job (gh)
  RuntimeError: Command docker exec -t 6e2cdcf1e606984603826c0e95a26216d56cca34f5836221d068806f715db0db /exec failed with exit code 2

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[kimishpatel]

looks largely good. Can you add some perf numbers on different backends

[mergennachin]

Putting back in the queue. See my minor previous comment

[suryasidd]

Putting back in the queue. See my minor previous comment

Separated the model definition and lowering.