diffusers/pr_13881/ko/_app/immutable/nodes/38.BFGVUIHP.js

import"../chunks/DsnmJJEf.js";import{i as b,h as u,C as f,H as M,a as d,E as y,s as w}from"../chunks/CFM6C53a.js";import{p as j,o as J,s as a,f as B,a as r,b as T,c,n as k}from"../chunks/CNc7KuUZ.js";import{D as Z}from"../chunks/BK2xlcGK.js";const v='{"title":"Text-guided depth-to-image 생성","local":"text-guided-depth-to-image-생성","sections":[],"depth":1}';var U=c('<meta name="hf:doc:metadata"/>'),W=c('<p></p> <!> <!> <!> <p><code>StableDiffusionDepth2ImgPipeline</code>을 사용하면 텍스트 프롬프트와 초기 이미지를 전달하여 새 이미지의 생성을 조절할 수 있습니다. 또한 이미지 구조를 보존하기 위해 <code>depth_map</code>을 전달할 수도 있습니다. <code>depth_map</code>이 제공되지 않으면 파이프라인은 통합된 <a href="https://github.com/isl-org/MiDaS" rel="nofollow">depth-estimation model</a>을 통해 자동으로 깊이를 예측합니다.</p> <p>먼저 <code>StableDiffusionDepth2ImgPipeline</code>의 인스턴스를 생성합니다:</p> <!> <p>이제 프롬프트를 파이프라인에 전달합니다. 특정 단어가 이미지 생성을 가이드 하는것을 방지하기 위해 <code>negative_prompt</code>를 전달할 수도 있습니다:</p> <!> <table><thead><tr><th>Input</th><th>Output</th></tr></thead><tbody><tr><td><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/coco-cats.png" width="500"/></td><td><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/depth2img-tigers.png" width="500"/></td></tr></tbody></table> <p>아래의 Spaces를 가지고 놀며 depth map이 있는 이미지와 없는 이미지의 차이가 있는지 확인해 보세요!</p> <iframe src="https://radames-stable-diffusion-depth2img.hf.space" frameborder="0" width="850" height="500"></iframe> <!> <p></p>',1);function G(g,m){j(m,!1),J(()=>{new URLSearchParams(window.location.search).get("fw")}),b();var e=W();u("1t2itkw",n=>{var p=U();w(p,"content",v),r(n,p)});var t=a(B(e),2);f(t,{containerStyle:"float: right; margin-left: 10px; display: inline-flex; position: relative; z-index: 10;"});var s=a(t,2);Z(s,{containerStyle:"float: right; margin-left: 10px; display: inline-flex; position: relative; z-index: 10;",options:[{label:"Mixed",value:"https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers_doc/ko/depth2img.ipynb"},{label:"PyTorch",value:"https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers_doc/ko/pytorch/depth2img.ipynb"},{label:"TensorFlow",value:"https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers_doc/ko/tensorflow/depth2img.ipynb"},{label:"Mixed",value:"https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/diffusers_doc/ko/depth2img.ipynb"},{label:"PyTorch",value:"https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/diffusers_doc/ko/pytorch/depth2img.ipynb"},{label:"TensorFlow",value:"https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/diffusers_doc/ko/tensorflow/depth2img.ipynb"}]});var o=a(s,2);M(o,{title:"Text-guided depth-to-image 생성",local:"text-guided-depth-to-image-생성",headingTag:"h1"});var l=a(o,6);d(l,{code:"aW1wb3J0JTIwdG9yY2glMEFpbXBvcnQlMjByZXF1ZXN0cyUwQWZyb20lMjBQSUwlMjBpbXBvcnQlMjBJbWFnZSUwQSUwQWZyb20lMjBkaWZmdXNlcnMlMjBpbXBvcnQlMjBTdGFibGVEaWZmdXNpb25EZXB0aDJJbWdQaXBlbGluZSUwQSUwQXBpcGUlMjAlM0QlMjBTdGFibGVEaWZmdXNpb25EZXB0aDJJbWdQaXBlbGluZS5mcm9tX3ByZXRyYWluZWQoJTBBJTIwJTIwJTIwJTIwJTIyc3RhYmlsaXR5YWklMkZzdGFibGUtZGlmZnVzaW9uLTItZGVwdGglMjIlMkMlMEElMjAlMjAlMjAlMjB0b3JjaF9kdHlwZSUzRHRvcmNoLmZsb2F0MTYlMkMlMEEpLnRvKCUyMmN1ZGElMjIp",highlighted:`<span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">import</span> requests
<span class="hljs-keyword">from</span> PIL <span class="hljs-keyword">import</span> Image

<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> StableDiffusionDepth2ImgPipeline

pipe = StableDiffusionDepth2ImgPipeline.from_pretrained(
    <span class="hljs-string">&quot;stabilityai/stable-diffusion-2-depth&quot;</span>,
    torch_dtype=torch.float16,
).to(<span class="hljs-string">&quot;cuda&quot;</span>)`,lang:"python",wrap:!1});var i=a(l,4);d(i,{code:"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",highlighted:`url = <span class="hljs-string">&quot;http://images.cocodataset.org/val2017/000000039769.jpg&quot;</span>
init_image = Image.<span class="hljs-built_in">open</span>(requests.get(url, stream=<span class="hljs-literal">True</span>).raw)
prompt = <span class="hljs-string">&quot;two tigers&quot;</span>
n_prompt = <span class="hljs-string">&quot;bad, deformed, ugly, bad anatomy&quot;</span>
image = pipe(prompt=prompt, image=init_image, negative_prompt=n_prompt, strength=<span class="hljs-number">0.7</span>).images[<span class="hljs-number">0</span>]
image`,lang:"python",wrap:!1});var h=a(i,8);y(h,{source:"https://github.com/huggingface/diffusers/blob/main/docs/source/ko/using-diffusers/depth2img.md"}),k(2),r(g,e),T()}export{G as component};