Bringing Nunchaku 4-bit Diffusion Inference to Diffusers

Nunchaku 4-bit diffusion inference is now natively integrated into Diffusers. The SVDQuant method runs transformer layers with 4-bit weights and activations, reducing VRAM usage by up to 50% while speeding up the denoising loop. Loading a checkpoint requires only from_pretrained() with no separate inference engine or local CUDA compilation.