AWPortrait-QW

AWPortrait-QW is based on the Qwen-Image. It's trained using a training set that better reflects Chinese facial features and aesthetics. It covers a wide range of genres, including indoor and outdoor portraits, fashion, and studio portraits, ensuring strong generalization. Compared to the original Qwen-Image, AWPortrait-QW delivers a more detailed and realistic rendering of skin texture.

Showcases

Trigger words

No trigger words are requireds. LoRA recommends a weight of 1.0.

Inference

import torch
from diffusers import QwenImagePipeline

pipe = QwenImagePipeline.from_pretrained("Qwen/Qwen-Image", torch_dtype=torch.bfloat16)
pipe.load_lora_weights("Shakker-Labs/AWPortrait-QW", weight_name="AWPortrait-QW_1.0.safetensors")
pipe.fuse_lora(lora_scale=1.0)
pipe.to("cuda")

prompt = "Black and white portrait of an Asian woman with dynamic hair movement,wearing a dark jacket against a light background."

image = pipe(
    prompt=prompt,
    negative_prompt="blurry, bad faces, bad hands, worst quality, low quality, jpeg artifacts",
    width=1328,
    height=1328,
    num_inference_steps=30,
    true_cfg_scale=4.0,
    generator=torch.Generator(device="cuda").manual_seed(0),
).images[0]
image.save(f"example.png")

Online Inference

You can also try this model on Liblib AI.

Acknowledgements

This model is trained by our copyrighted users DynamicWang. We release this model under permissions.