prithivMLmods/Qwen-Image-Edit-Rapid-AIO-V21

Diffusers-compatible transformer weights extracted from Phr00t/Qwen-Image-Edit-Rapid-AIO-NSFW-V21 for 4-step accelerated Qwen Image Edit inference.

Tested some sample inferences for Phr00t/Qwen-Image-Edit-Rapid-AIO v21 @ 4 steps. Still, v19 is the best in overall performance. v21 performs better @ 6–8 steps or even higher.

Quick Start with Diffusers🧨

pip install -U torch transformers diffusers

import torch
from diffusers.models import QwenImageTransformer2DModel
from diffusers import QwenImageEditPlusPipeline
from diffusers.utils import load_image

transformer = QwenImageTransformer2DModel.from_pretrained(
    "prithivMLmods/Qwen-Image-Edit-Rapid-AIO-V21",
    torch_dtype=torch.bfloat16
)

pipeline = QwenImageEditPlusPipeline.from_pretrained(
    "Qwen/Qwen-Image-Edit-2511",
    transformer=transformer,
    torch_dtype=torch.bfloat16
)
pipeline.to("cuda")

image1 = load_image("grumpycat.png")
prompt = "turn the cat into an orange cat"

inputs = {
    "image": [image1],
    "prompt": prompt,
    "generator": torch.manual_seed(42),
    "true_cfg_scale": 1.0,
    "negative_prompt": " ",
    "num_inference_steps": 4,
    "guidance_scale": 1.0,
    "num_images_per_prompt": 1,
}

output = pipeline(**inputs)
output_image = output.images[0]
output_image.save("output_image_edit_plus.png")