[WIP] Refactor CPU Offloading Backend Pattern

[yuanheng-zhao]

PLEASE FILL IN THE PR DESCRIPTION HERE ENSURING ALL CHECKLIST ITEMS (AT THE BOTTOM) HAVE BEEN CONSIDERED.

Purpose

Closes #1125

Test Plan

T2V, I2V offline examples with Wan pipeline

Test Result

To be added

Layer-wise

Logs

[Stage-0] INFO 02-05 14:19:12 [diffusion_model_runner.py:109] Model runner: Model loaded successfully.
[Stage-0] INFO 02-05 14:19:12 [layerwise_backend.py:323] Applying layer-wise offloading on ['transformer', 'transformer_2']
[Stage-0] INFO 02-05 14:19:12 [layerwise_backend.py:328] Applying hooks on transformer (WanTransformer3DModel)
[Stage-0] INFO 02-05 14:19:38 [layerwise_backend.py:358] Layer-wise offloading enabled on 40 layers (blocks), with 1 kept on device
[Stage-0] INFO 02-05 14:19:38 [layerwise_backend.py:328] Applying hooks on transformer_2 (WanTransformer3DModel)
[Stage-0] INFO 02-05 14:20:02 [layerwise_backend.py:358] Layer-wise offloading enabled on 40 layers (blocks), with 1 kept on device
[Stage-0] INFO 02-05 14:20:02 [diffusion_model_runner.py:125] Model runner: Model compiled with torch.compile.

---

Essential Elements of an Effective PR Description Checklist

• The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
• The test plan, such as providing test command.
• The test results, such as pasting the results comparison before and after, or e2e results
• (Optional) The necessary documentation update, such as updating supported_models.md and examples for a new model.
• (Optional) Release notes update. If your change is user facing, please update the release notes draft.

BEFORE SUBMITTING, PLEASE READ https://github.com/vllm-project/vllm-omni/blob/main/CONTRIBUTING.md (anything written below this line will be removed by GitHub Actions)