Support for ByteDance's Lance Model (Omni Multi-modal)

### Feature Idea

With the rise of efficient, multi-modal omni models, there is currently no official support in ComfyUI for unified text, audio, and video generation/understanding within a single compact model.

It would be amazing to have official ComfyUI support (nodes/loaders) for ByteDance's new Lance model.

Repository: [bytedance/Lance](https://github.com/bytedance/Lance)

Hugging Face: [bytedance-research/Lance](https://huggingface.co/bytedance-research/Lance)

Why this would be a great addition:

Efficiency: At only 3B parameters, it is highly accessible for consumer hardware and local running.

Versatility: It natively supports text, audio, and video (both generation and understanding), which opens up massive possibilities for complex, multi-modal ComfyUI workflows.

Thank you to the ComfyUI team for all your incredible hard work and for sharing this amazing tool with the community!

### Existing Solutions

_No response_

### Other

_No response_