免费开放商用！Stability AI推轻量级AI绘画利器 Stable Diffusion 3.5 Medium模型

Stability AI再次突破技术壁垒，推出全新Stable Diffusion3.5Medium模型。这款面向大众的AI绘画工具不仅完全免费开放商用，更重要的是实现了高性能与普及性的完美平衡。

这款采用多模态扩散变换器（MMDiT-X）架构的模型，以25亿参数的精简设计，巧妙解决了普通用户的硬件门槛问题。仅需9.9GB显存，便能在大多数消费级显卡上流畅运行，真正实现了"人人可用"的愿景。

在技术创新方面，该模型整合了三种预训练文本编码器，并引入QK标准化技术提升训练稳定性。特别值得一提的是，其前12个变换层中的双重注意力模块设计，让模型在图像质量、排版效果和复杂提示理解等方面都有显著提升。

模型的训练过程融合了合成数据与精选公共数据，采用渐进式分辨率提升的混合训练策略，确保了生成图像的多样性和质量。与同类中型模型相比，它在图像生成效果和处理速度上都展现出明显优势。

不过，用户在使用过程中需要注意一些细节:过长的提示词可能导致图像边缘出现瑕疵;建议使用跳层指导采样方式来优化图像的结构完整性;同时要注意，由于训练数据分布的差异，相同提示词可能会产生不同的创作效果。

这款模型的发布，不仅为个人创作者和初创企业提供了便捷的AI创作工具，更体现了Stability AI推动AI技术普及化的决心。无论是用于艺术创作还是教育开发，它都将为更广泛的用户群体带来AI创作的可能性。

模型下载地址：https://huggingface.co/stabilityai/stable-diffusion-3.5-medium

架构

Stable Diffusion 3.5 Medium 是一款改进型多模态扩散转换器（MMDiT-X）文本到图像模型，在图像质量、排版、复杂提示理解和资源效率方面的性能都有所提高。

├── text_encoders/  
│   ├── README.md
│   ├── clip_g.safetensors
│   ├── clip_l.safetensors
│   ├── t5xxl_fp16.safetensors
│   └── t5xxl_fp8_e4m3fn.safetensors
│
├── README.md
├── LICENSE
├── sd3.5_medium.safetensors
├── SD3.5M_example_workflow.json
├── SD3.5M_SLG_example_workflow.json
├── SD3.5L_plus_SD3.5M_upscaling_example_workflow.json
└── sd3_medium_demo.jpg

** File structure below is for diffusers integration**
├── scheduler/
├── text_encoder/
├── text_encoder_2/
├── text_encoder_3/
├── tokenizer/
├── tokenizer_2/
├── tokenizer_3/
├── transformer/
├── vae/
└── model_index.json

Diffusers

pip install -U diffusers

import torch
from diffusers import StableDiffusion3Pipeline

pipe = StableDiffusion3Pipeline.from_pretrained("stabilityai/stable-diffusion-3.5-medium", torch_dtype=torch.bfloat16)
pipe = pipe.to("cuda")

image = pipe(
    "A capybara holding a sign that reads Hello World",
    num_inference_steps=40,
    guidance_scale=4.5,
).images[0]
image.save("capybara.png")

使用扩散器量化模型

减少 VRAM 使用量，让模型适合 🤏 VRAM GPU

pip install bitsandbytes

from diffusers import BitsAndBytesConfig, SD3Transformer2DModel
from diffusers import StableDiffusion3Pipeline
import torch

model_id = "stabilityai/stable-diffusion-3.5-medium"

nf4_config = BitsAndBytesConfig(
    load_in_4bit=True,
    bnb_4bit_quant_type="nf4",
    bnb_4bit_compute_dtype=torch.bfloat16
)
model_nf4 = SD3Transformer2DModel.from_pretrained(
    model_id,
    subfolder="transformer",
    quantization_config=nf4_config,
    torch_dtype=torch.bfloat16
)

pipeline = StableDiffusion3Pipeline.from_pretrained(
    model_id, 
    transformer=model_nf4,
    torch_dtype=torch.bfloat16
)
pipeline.enable_model_cpu_offload()

prompt = "A whimsical and creative image depicting a hybrid creature that is a mix of a waffle and a hippopotamus, basking in a river of melted butter amidst a breakfast-themed landscape. It features the distinctive, bulky body shape of a hippo. However, instead of the usual grey skin, the creature's body resembles a golden-brown, crispy waffle fresh off the griddle. The skin is textured with the familiar grid pattern of a waffle, each square filled with a glistening sheen of syrup. The environment combines the natural habitat of a hippo with elements of a breakfast table setting, a river of warm, melted butter, with oversized utensils or plates peeking out from the lush, pancake-like foliage in the background, a towering pepper mill standing in for a tree.  As the sun rises in this fantastical world, it casts a warm, buttery glow over the scene. The creature, content in its butter river, lets out a yawn. Nearby, a flock of birds take flight"

image = pipeline(
    prompt=prompt,
    num_inference_steps=40,
    guidance_scale=4.5,
    max_sequence_length=512,
).images[0]
image.save("whimsical.png")