[!NOTE]
This is a GGUF quantized version of Qwen-Image-Layered.
unsloth/Qwen-Image-Layered-GGUF uses Unsloth Dynamic 2.0 methodology for SOTA performance. Important layers are upcasted to higher precision.
We are excited to introduce Qwen-Image-Layered, a model capable of decomposing an image into multiple RGBA layers. This layered representation unlocks inherent editability: each layer can be independently manipulated without affecting other content. Meanwhile, such a layered representation naturally supports high-fidelity elementary operations-such as resizing, reposition, and recoloring. By physically isolating semantic or structural components into distinct layers, our approach enables high-fidelity and consistent editing.
Quick Start
Make sure your transformers>=4.51.3 (Supporting Qwen2.5-VL)
1from diffusers import QwenImageLayeredPipeline
2import torch
3from PIL import Image
45pipeline = QwenImageLayeredPipeline.from_pretrained("Qwen/Qwen-Image-Layered")6pipeline = pipeline.to("cuda", torch.bfloat16)7pipeline.set_progress_bar_config(disable=None)89image = Image.open("asserts/test_images/1.png").convert("RGBA")10inputs ={11"image": image,12"generator": torch.Generator(device='cuda').manual_seed(777),13"true_cfg_scale":4.0,14"negative_prompt":" ",15"num_inference_steps":50,16"num_images_per_prompt":1,17"layers":4,18"resolution":640,# Using different bucket (640, 1024) to determine the resolution. For this version, 640 is recommended19"cfg_normalize":True,# Whether enable cfg normalization.20"use_en_prompt":True,# Automatic caption language if user does not provide caption21}2223with torch.inference_mode():24 output = pipeline(**inputs)25 output_image = output.images[0]2627for i, image inenumerate(output_image):28 image.save(f"{i}.png")
Showcase
Layered Decomposition in Application
Given an image, Qwen-Image-Layered can decompose it into several RGBA layers:
Example Image
After decomposition, edits are applied exclusively to the target layer, physically isolating it from the rest of the content, and thereby fundamentally ensuring consistency across edits.
For example, we can recolor the first layer and keep all other content untouched:
Example Image
We can also replace the second layer from a girl to a boy (The target layer is edited using Qwen-Image-Edit):
Example Image
Here, we revise the text to "Qwen-Image" (The target layer is edited using Qwen-Image-Edit):
Example Image
Furthermore, the layered structure naturally supports elemetary operations. For example, we can delete unwanted objects cleanly:
Example Image
We can also resize an object without distortion:
Example Image
After layer decomposition, we can move objects freely within the canvas:
Example Image
Flexible and Iterative Decomposition
Qwen-Image-Layered is not limited to a fixed number of layers. The model supports variable-layer decomposition. For example, we can decompose an image into either 3 or 8 layers as needed:
Example Image
Moreover, decomposition can be applied recursively: any layer can itself be further decomposed, enabling infinite decomposition.
Example Image
License Agreement
Qwen-Image-Layered is licensed under Apache 2.0.
Citation
We kindly encourage citation of our work if you find it useful.
