diffusers/pr_13881/ko/_app/immutable/nodes/52.fKjf-L5q.js

import"../chunks/DsnmJJEf.js";import{i as U,h as f,C as w,H as j,a,E as _,s as J}from"../chunks/CFM6C53a.js";import{p as v,o as k,s,f as W,a as b,b as T,c as u,n as C}from"../chunks/CNc7KuUZ.js";import{D as B}from"../chunks/BK2xlcGK.js";const Z='{"title":"Textual inversion","local":"textual-inversion","sections":[],"depth":1}';var G=u('<meta name="hf:doc:metadata"/>'),Q=u('<p></p> <!> <!> <!> <p><code>StableDiffusionPipeline</code>은  textual-inversion을 지원하는데, 이는 몇 개의 샘플 이미지만으로 stable diffusion과 같은 모델이 새로운 컨셉을 학습할 수 있도록 하는 기법입니다. 이를 통해 생성된 이미지를 더 잘 제어하고 특정 컨셉에 맞게 모델을 조정할 수 있습니다. 커뮤니티에서 만들어진 컨셉들의 컬렉션은 <a href="https://huggingface.co/spaces/sd-concepts-library/stable-diffusion-conceptualizer" rel="nofollow">Stable Diffusion Conceptualizer</a>를 통해 빠르게 사용해볼 수 있습니다.</p> <p>이 가이드에서는 Stable Diffusion Conceptualizer에서 사전학습한 컨셉을 사용하여 textual-inversion으로 추론을 실행하는 방법을 보여드립니다. textual-inversion으로 모델에 새로운 컨셉을 학습시키는 데 관심이 있으시다면, <a href="./training/text_inversion">Textual Inversion</a> 훈련 가이드를 참조하세요.</p> <p>Hugging Face 계정으로 로그인하세요:</p> <!> <p>필요한 라이브러리를 불러오고 생성된 이미지를 시각화하기 위한 도우미 함수 <code>image_grid</code>를 만듭니다:</p> <!> <p>Stable Diffusion과 <a href="https://huggingface.co/spaces/sd-concepts-library/stable-diffusion-conceptualizer" rel="nofollow">Stable Diffusion Conceptualizer</a>에서 사전학습된 컨셉을 선택합니다:</p> <!> <p>이제 파이프라인을 로드하고 사전학습된 컨셉을 파이프라인에 전달할 수 있습니다:</p> <!> <p>특별한 placeholder token ’<code>&lt;cat-toy&gt;</code>‘를 사용하여 사전학습된 컨셉으로 프롬프트를 만들고, 생성할 샘플의 수와 이미지 행의 수를 선택합니다:</p> <!> <p>그런 다음 파이프라인을 실행하고, 생성된 이미지들을 저장합니다. 그리고 처음에 만들었던 도우미 함수 <code>image_grid</code>를 사용하여 생성 결과들을 시각화합니다. 이 때 <code>num_inference_steps</code>와 <code>guidance_scale</code>과 같은 매개 변수들을 조정하여, 이것들이 이미지 품질에 어떠한 영향을 미치는지를 자유롭게 확인해보시기 바랍니다.</p> <!> <div class="flex justify-center"><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/textual_inversion_inference.png"/></div> <!> <p></p>',1);function I(g,m){v(m,!1),k(()=>{new URLSearchParams(window.location.search).get("fw")}),U();var l=Q();f("1mqg1pd",y=>{var M=G();J(M,"content",Z),b(y,M)});var e=s(W(l),2);w(e,{containerStyle:"float: right; margin-left: 10px; display: inline-flex; position: relative; z-index: 10;"});var n=s(e,2);B(n,{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/textual_inversion_inference.ipynb"},{label:"PyTorch",value:"https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers_doc/ko/pytorch/textual_inversion_inference.ipynb"},{label:"TensorFlow",value:"https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers_doc/ko/tensorflow/textual_inversion_inference.ipynb"},{label:"Mixed",value:"https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/diffusers_doc/ko/textual_inversion_inference.ipynb"},{label:"PyTorch",value:"https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/diffusers_doc/ko/pytorch/textual_inversion_inference.ipynb"},{label:"TensorFlow",value:"https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/diffusers_doc/ko/tensorflow/textual_inversion_inference.ipynb"}]});var o=s(n,2);j(o,{title:"Textual inversion",local:"textual-inversion",headingTag:"h1"});var i=s(o,8);a(i,{code:"ZnJvbSUyMGh1Z2dpbmdmYWNlX2h1YiUyMGltcG9ydCUyMG5vdGVib29rX2xvZ2luJTBBJTBBbm90ZWJvb2tfbG9naW4oKQ==",highlighted:`<span class="hljs-keyword">from</span> huggingface_hub <span class="hljs-keyword">import</span> notebook_login

notebook_login()`,lang:"py",wrap:!1});var t=s(i,4);a(t,{code:"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",highlighted:`<span class="hljs-keyword">import</span> os
<span class="hljs-keyword">import</span> torch

<span class="hljs-keyword">import</span> PIL
<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> StableDiffusionPipeline
<span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> CLIPImageProcessor, CLIPTextModel, CLIPTokenizer


<span class="hljs-keyword">def</span> <span class="hljs-title function_">image_grid</span>(<span class="hljs-params">imgs, rows, cols</span>):
    <span class="hljs-keyword">assert</span> <span class="hljs-built_in">len</span>(imgs) == rows * cols

    w, h = imgs[<span class="hljs-number">0</span>].size
    grid = Image.new(<span class="hljs-string">&quot;RGB&quot;</span>, size=(cols * w, rows * h))
    grid_w, grid_h = grid.size

    <span class="hljs-keyword">for</span> i, img <span class="hljs-keyword">in</span> <span class="hljs-built_in">enumerate</span>(imgs):
        grid.paste(img, box=(i % cols * w, i // cols * h))
    <span class="hljs-keyword">return</span> grid`,lang:"py",wrap:!1});var r=s(t,4);a(r,{code:"cHJldHJhaW5lZF9tb2RlbF9uYW1lX29yX3BhdGglMjAlM0QlMjAlMjJzdGFibGUtZGlmZnVzaW9uLXYxLTUlMkZzdGFibGUtZGlmZnVzaW9uLXYxLTUlMjIlMEFyZXBvX2lkX2VtYmVkcyUyMCUzRCUyMCUyMnNkLWNvbmNlcHRzLWxpYnJhcnklMkZjYXQtdG95JTIy",highlighted:`pretrained_model_name_or_path = <span class="hljs-string">&quot;stable-diffusion-v1-5/stable-diffusion-v1-5&quot;</span>
repo_id_embeds = <span class="hljs-string">&quot;sd-concepts-library/cat-toy&quot;</span>`,lang:"py",wrap:!1});var p=s(r,4);a(p,{code:"cGlwZWxpbmUlMjAlM0QlMjBTdGFibGVEaWZmdXNpb25QaXBlbGluZS5mcm9tX3ByZXRyYWluZWQocHJldHJhaW5lZF9tb2RlbF9uYW1lX29yX3BhdGglMkMlMjB0b3JjaF9kdHlwZSUzRHRvcmNoLmZsb2F0MTYpLnRvKCUyMmN1ZGElMjIpJTBBJTBBcGlwZWxpbmUubG9hZF90ZXh0dWFsX2ludmVyc2lvbihyZXBvX2lkX2VtYmVkcyk=",highlighted:`pipeline = StableDiffusionPipeline.from_pretrained(pretrained_model_name_or_path, torch_dtype=torch.float16).to(<span class="hljs-string">&quot;cuda&quot;</span>)

pipeline.load_textual_inversion(repo_id_embeds)`,lang:"py",wrap:!1});var c=s(p,4);a(c,{code:"cHJvbXB0JTIwJTNEJTIwJTIyYSUyMGdyYWZpdHRpJTIwaW4lMjBhJTIwZmF2ZWxhJTIwd2FsbCUyMHdpdGglMjBhJTIwJTNDY2F0LXRveSUzRSUyMG9uJTIwaXQlMjIlMEElMEFudW1fc2FtcGxlcyUyMCUzRCUyMDIlMEFudW1fcm93cyUyMCUzRCUyMDI=",highlighted:`prompt = <span class="hljs-string">&quot;a grafitti in a favela wall with a &lt;cat-toy&gt; on it&quot;</span>

num_samples = <span class="hljs-number">2</span>
num_rows = <span class="hljs-number">2</span>`,lang:"py",wrap:!1});var d=s(c,4);a(d,{code:"YWxsX2ltYWdlcyUyMCUzRCUyMCU1QiU1RCUwQWZvciUyMF8lMjBpbiUyMHJhbmdlKG51bV9yb3dzKSUzQSUwQSUyMCUyMCUyMCUyMGltYWdlcyUyMCUzRCUyMHBpcGUocHJvbXB0JTJDJTIwbnVtX2ltYWdlc19wZXJfcHJvbXB0JTNEbnVtX3NhbXBsZXMlMkMlMjBudW1faW5mZXJlbmNlX3N0ZXBzJTNENTAlMkMlMjBndWlkYW5jZV9zY2FsZSUzRDcuNSkuaW1hZ2VzJTBBJTIwJTIwJTIwJTIwYWxsX2ltYWdlcy5leHRlbmQoaW1hZ2VzKSUwQSUwQWdyaWQlMjAlM0QlMjBpbWFnZV9ncmlkKGFsbF9pbWFnZXMlMkMlMjBudW1fc2FtcGxlcyUyQyUyMG51bV9yb3dzKSUwQWdyaWQ=",highlighted:`all_images = []
<span class="hljs-keyword">for</span> _ <span class="hljs-keyword">in</span> <span class="hljs-built_in">range</span>(num_rows):
    images = pipe(prompt, num_images_per_prompt=num_samples, num_inference_steps=<span class="hljs-number">50</span>, guidance_scale=<span class="hljs-number">7.5</span>).images
    all_images.extend(images)

grid = image_grid(all_images, num_samples, num_rows)
grid`,lang:"py",wrap:!1});var h=s(d,4);_(h,{source:"https://github.com/huggingface/diffusers/blob/main/docs/source/ko/using-diffusers/textual_inversion_inference.md"}),C(2),b(g,l),T()}export{I as component};