SD3.5 LoRA Linear Red Light

🚀 SD3.5-LoRA-Linear-Red-Light

這是一個基於StableDiffusion 3.5的LoRA模型，可在圖像生成中添加線性紅光效果。

🚀 快速開始

觸發詞

你應該使用 Linear red light 來觸發圖像生成。

推理示例

以下是使用該模型進行推理的Python代碼示例：

import torch
from diffusers import StableDiffusion3Pipeline # pip install diffusers>=0.31.0

pipe = StableDiffusion3Pipeline.from_pretrained("stabilityai/stable-diffusion-3.5-large", torch_dtype=torch.bfloat16)
pipe.load_lora_weights("Shakker-Labs/SD3.5-LoRA-Linear-Red-Light", weight_name="SD35-lora-Linear-Red-Light.safetensors")
pipe.fuse_lora(lora_scale=1.0)
pipe.to("cuda")

prompt = "a spider-man, Linear red light"
negative_prompt = "(lowres, low quality, worst quality)"

image = pipe(prompt=prompt,
             negative_prompt=negative_prompt,
             num_inference_steps=24, 
             guidance_scale=4.0,
             width=960, height=1280,
            ).images[0]
image.save(f"toy_example.jpg")

💻 使用示例

基礎用法

import torch
from diffusers import StableDiffusion3Pipeline # pip install diffusers>=0.31.0

pipe = StableDiffusion3Pipeline.from_pretrained("stabilityai/stable-diffusion-3.5-large", torch_dtype=torch.bfloat16)
pipe.load_lora_weights("Shakker-Labs/SD3.5-LoRA-Linear-Red-Light", weight_name="SD35-lora-Linear-Red-Light.safetensors")
pipe.fuse_lora(lora_scale=1.0)
pipe.to("cuda")

prompt = "a spider-man, Linear red light"
negative_prompt = "(lowres, low quality, worst quality)"

image = pipe(prompt=prompt,
             negative_prompt=negative_prompt,
             num_inference_steps=24, 
             guidance_scale=4.0,
             width=960, height=1280,
            ).images[0]
image.save(f"toy_example.jpg")

高級用法

# 可以根據需要調整推理參數，如推理步數、引導比例、圖像尺寸等，以獲得不同效果的圖像。
import torch
from diffusers import StableDiffusion3Pipeline # pip install diffusers>=0.31.0

pipe = StableDiffusion3Pipeline.from_pretrained("stabilityai/stable-diffusion-3.5-large", torch_dtype=torch.bfloat16)
pipe.load_lora_weights("Shakker-Labs/SD3.5-LoRA-Linear-Red-Light", weight_name="SD35-lora-Linear-Red-Light.safetensors")
pipe.fuse_lora(lora_scale=1.0)
pipe.to("cuda")

prompt = "a beautiful princess, Linear red light"
negative_prompt = "(lowres, low quality, worst quality)"

# 調整推理步數和引導比例
image = pipe(prompt=prompt,
             negative_prompt=negative_prompt,
             num_inference_steps=30, 
             guidance_scale=5.0,
             width=1024, height=1024,
            ).images[0]
image.save(f"advanced_example.jpg")

📄 許可證

本模型使用的許可證為 stabilityai-ai-community，你可以通過以下鏈接查看許可證詳情：https://huggingface.co/stabilityai/stable-diffusion-3-medium/blob/main/LICENSE.md

信息表格