Please add support for GGUF in Lumina2 pipeline

[nitinmukesh]

Is your feature request related to a problem? Please describe.
GGUF is already available, please add support in pipeline
https://huggingface.co/calcuis/lumina-gguf/tree/main

Describe the solution you'd like.

import torch
from diffusers import Lumina2Text2ImgPipeline, Lumina2Transformer2DModel
bfl_repo = "Alpha-VLLM/Lumina-Image-2.0"
dtype = torch.bfloat16
transformer_path = f"https://huggingface.co/calcuis/lumina-gguf/blob/main/lumina2-q8_0.gguf"
transformer = Lumina2Transformer2DModel.from_single_file(
	transformer_path,
	quantization_config=GGUFQuantizationConfig(compute_dtype=torch.bfloat16),
	torch_dtype=dtype,
	config=bfl_repo,
	subfolder="transformer"
)

pipe = Lumina2Text2ImgPipeline.from_pretrained(
	bfl_repo,
	transformer=transformer,
	torch_dtype=dtype,
)
pipe.enable_model_cpu_offload()
pipe.vae.enable_slicing()
pipe.vae.enable_tiling()
inference_params = {
	"prompt": "Portrait of a young woman in a Victorian-era outfit with brass goggles and leather straps. Background shows an industrial revolution cityscape with smoky skies and tall, metal structures",
	"height": 1024,
	"width": 576,
	"guidance_scale": 4.0,
	"num_inference_steps": 30,
	"generator": torch.Generator(device="cpu").manual_seed(0),
}
image = pipe(**inference_params).images[0]
image.save(output_path)

Describe alternatives you've considered.
BnB int4 / int8 works, with GGUF we may achieve further memory reduction.

Additional context.
(venv) C:\aiOWN\diffuser_webui>python lumina2_gguf.py
Traceback (most recent call last):
File "C:\aiOWN\diffuser_webui\lumina2_gguf.py", line 6, in
transformer = Lumina2Transformer2DModel.from_single_file(
AttributeError: type object 'Lumina2Transformer2DModel' has no attribute 'from_single_file'

@zhuole1025

[nitinmukesh]

from_single_file support will also help use bf16/fp16
https://huggingface.co/Comfy-Org/Lumina_Image_2.0_Repackaged/tree/main/split_files

[DN6]

Closing since #10781 was merged.