[enhancement]: Add support for SVDQuant

### Is there an existing issue for this?

- [x] I have searched the existing issues

### Contact Details

_No response_

### What should this feature add?

Hi,

SVDQuant allows to use FLUX on low end hardware by reducing memory usage and increasing performance. For example, with my RTX 3050 (8GB), the iteration time goes from 6 seconds per iteration to 1.5 seconds, eliminating the need to keep using SDXL (at least in my case). At the moment, the project is completely usable even though they are adding new features and fixing some minor bugs. In fact, I am using it within ComfyUI, but I would really love to be able to use it within InvokeAI, because the interface is so much better.

**References**:
https://github.com/mit-han-lab/nunchaku/tree/main
https://github.com/mit-han-lab/ComfyUI-nunchaku

Best regards,
Robert.

### Alternatives

_No response_

### Additional Content

_No response_