starsfriday Qwen-Image-Edit-2511-Trendy_figurines LoRA

Model Card for Model ID

This is a Qwen-Image-Edit-2511 LoRA trained on Qwen/Qwen-Image-Edit-2511. It is a general-purpose image editing LoRA for transforming ordinary input images into a trendy figurine / blind-box toy style with Pop Mart-inspired aesthetics, cute chibi proportions, smooth PVC material, delicate gloss, soft colors, and high-detail 3D rendering. The model can be used with a broad range of common images, including people, portraits, full-body characters, animals, pets, objects, products, indoor scenes, outdoor scenes, lifestyle photos, and mixed real-world compositions. It focuses on preserving the main subject identity and recognizable structure while restyling the image into a collectible toy / 3D figurine look. For use in ComfyUI or any Qwen-Image-Edit-2511 compatible LoRA inference pipeline.

ComfyUI Workflow

This LoRA works with a modified version of Comfy's Qwen/Qwen-Image-Edit-2511 workflow. The main modification is adding a Qwen-Image-Edit-2511 LoRA node connected to the base model.

Put qwen_image_edit_2511_trendy_figurines.safetensors into ComfyUI/models/loras/, import the linked workflow in ComfyUI, then select this LoRA weight in the workflow.

See the Downloads section below for the model weights and workflow file.

Direct Use

from diffusers import QwenImageEditPipeline
import torch
from PIL import Image

# Load the Qwen-Image-Edit-2511 pipeline.
pipeline = QwenImageEditPipeline.from_pretrained(
    "Qwen/Qwen-Image-Edit-2511",
    torch_dtype=torch.bfloat16,
)
pipeline.to("cuda")

# Load trained LoRA weights for trendy figurine image editing.
pipeline.load_lora_weights(
    "starsfriday/Qwen-Image-Edit-2511-Trendy_figurines",
    weight_name="qwen_image_edit_2511_trendy_figurines.safetensors",
)

# Load input image.
image = Image.open("./input.png").convert("RGB")

# Define the editing prompt.
prompt = "把整个图片的内容变成潮玩手办风格，风格： 盲盒玩具风，泡泡玛特审美，Q版风格，材质： 光滑PVC材质，细腻光泽，细节： 精致细节，柔和配色，渲染： 3D渲染，Octane渲染，高细节，8k超高清。"

# Generate edited image.
inputs = {
    "image": image,
    "prompt": prompt,
    "generator": torch.manual_seed(12345),
    "true_cfg_scale": 3.0,
    "negative_prompt": " ",
    "num_inference_steps": 25,
}

with torch.inference_mode():
    output = pipeline(**inputs)
    output_image = output.images[0]
    output_image.save("edited_image.png")

Trigger phrase

把整个图片的内容变成潮玩手办风格，风格： 盲盒玩具风，泡泡玛特审美，Q版风格，材质： 光滑PVC材质，细腻光泽，细节： 精致细节，柔和配色，渲染： 3D渲染，Octane渲染，高细节，8k超高清。

There is no fixed trigger word. The prompt can be adjusted according to the input image, subject, pose, clothing, scene, lighting, and background.

Suggested settings

Resolution: 1024 x 1024
Sampler: flowmatch
Sampling steps: 25
Guidance scale: 3
LoRA strength: start from 0.8 to 1.0

Download model

Weights for this model are available in Safetensors format.

Download model

Download LoRA weights

Download ComfyUI workflow

Contact

Email: z@vvicat.com

Training at JoyFox Lab

This model family was trained and optimized by JoyFox Lab (Chengdu Xuanhu Technology Co., Ltd.).