UPDATE images

Kwai-Kolors · Jul 5, 2024 · 1b35b92 · 1b35b92
1 parent 7707d88
commit 1b35b92
Show file tree

Hide file tree

Showing 29 changed files with 79 additions and 1,958 deletions.
diff --git a/images/Pope-Edouard-de-Beaumont-1844.jpg b/images/Pope-Edouard-de-Beaumont-1844.jpg
diff --git a/images/Victor_Hugo-Hunchback.jpg b/images/Victor_Hugo-Hunchback.jpg
diff --git a/images/cvpr2024_learning/cvpr2024_learning_1.png b/images/cvpr2024_learning/cvpr2024_learning_1.png
diff --git a/images/esmeralda.jpg b/images/esmeralda.jpg
diff --git a/images/kolors-3b/background_55.png b/images/kolors-3b/background_55.png
diff --git a/images/post-2.png b/images/post-2.png
diff --git a/images/post-2/cn_all.png b/images/post-2/cn_all.png
diff --git a/images/post-2/fz_all.png b/images/post-2/fz_all.png
diff --git a/images/post-2/post-2.png b/images/post-2/post-2.png
diff --git a/images/post-2/wz_all.png b/images/post-2/wz_all.png
diff --git a/images/post-2/zl8.png b/images/post-2/zl8.png
diff --git a/index.html b/index.html
@@ -184,18 +184,18 @@ <h1 class="flex-none">
 
         <div class="pr3-ns mb4 mb0-ns w-100 w-40-ns">
           <a href="/post/post-2/" class="db grow">
-            <img src="https://Kwai-Kolors.github.io/images/post-2.png" class="img" alt="image from Kolors: Unet is enough for diffusion (arXiv-2024)">
+            <img src="https://Kwai-Kolors.github.io/images/post-2.png" class="img" alt="image from Kolors: Effective Training of Diffusion Model for Photorealistic Text-to-Image Synthesis (arXiv-2024)">
           </a>
         </div>
 
       <div class="blah w-100 w-60-ns pl3-ns">
         <h1 class="f3 fw1 athelas mt0 lh-title">
           <a href="/post/post-2/" class="color-inherit dim link">
-            Kolors: Unet is enough for diffusion (arXiv-2024)
+            Kolors: Effective Training of Diffusion Model for Photorealistic Text-to-Image Synthesis (arXiv-2024)
             </a>
         </h1>
         <div class="f6 f5-l lh-copy nested-copy-line-height nested-links">
-          We have open-sourced the <strong>Kolors</strong> large model, which is a large-scale text-to-image generation model based on latent diffusion. The current open-source model has 2.7 billion parameters. Kolors is trained on billions of text-image pairs and demonstrates significant advantages in visual quality, complex semantic understanding, and even text generation (Chinese and English characters) compared to both open-source and closed-source models. Additionally, Kolors supports both Chinese and English, making it more competitive in understanding Chinese-specific content. The model link is provided in the table below; please click to learn more.
+          Kolors is a large-scale text-to-image generation model based on latent diffusion, developed by the Kuaishou Kolors team. Trained on billions of text-image pairs, Kolors exhibits significant advantages over both open-source and proprietary models in visual quality, complex semantic accuracy, and text rendering for both Chinese and English characters. Furthermore, Kolors supports both Chinese and English inputs, demonstrating strong performance in understanding and generating Chinese-specific content. For more details, please refer to this <strong><a href="https://kolors.kuaishou.com/">technical report</a></strong>.
         </div>
           <a href="/post/post-2/" class="ba b--moon-gray bg-light-gray br2 color-inherit dib f7 hover-bg-moon-gray link mt2 ph2 pv1">read more</a>
 

diff --git a/index.xml b/index.xml
@@ -9,11 +9,11 @@
     <lastBuildDate>Fri, 21 Jun 2024 11:00:59 -0400</lastBuildDate>
     <atom:link href="https://Kwai-Kolors.github.io/index.xml" rel="self" type="application/rss+xml" />
     <item>
-      <title>Kolors: Unet is enough for diffusion (arXiv-2024)</title>
+      <title>Kolors: Effective Training of Diffusion Model for Photorealistic Text-to-Image Synthesis (arXiv-2024)</title>
       <link>https://Kwai-Kolors.github.io/post/post-2/</link>
       <pubDate>Fri, 21 Jun 2024 11:00:59 -0400</pubDate>
       <guid>https://Kwai-Kolors.github.io/post/post-2/</guid>
-      <description>We have open-sourced the &lt;strong&gt;Kolors&lt;/strong&gt; large model, which is a large-scale text-to-image generation model based on latent diffusion. The current open-source model has 2.7 billion parameters. Kolors is trained on billions of text-image pairs and demonstrates significant advantages in visual quality, complex semantic understanding, and even text generation (Chinese and English characters) compared to both open-source and closed-source models. Additionally, Kolors supports both Chinese and English, making it more competitive in understanding Chinese-specific content. The model link is provided in the table below; please click to learn more.</description>
+      <description>Kolors is a large-scale text-to-image generation model based on latent diffusion, developed by the Kuaishou Kolors team. Trained on billions of text-image pairs, Kolors exhibits significant advantages over both open-source and proprietary models in visual quality, complex semantic accuracy, and text rendering for both Chinese and English characters. Furthermore, Kolors supports both Chinese and English inputs, demonstrating strong performance in understanding and generating Chinese-specific content. For more details, please refer to this &lt;strong&gt;&lt;a href=&#34;https://kolors.kuaishou.com/&#34;&gt;technical report&lt;/a&gt;&lt;/strong&gt;.</description>
     </item>
     <item>
       <title>Learning Multi-dimensional Human Preference for Text-to-Image Generation (CVPR-2024)</title>

diff --git a/post/chapter-1/index.html b/post/chapter-1/index.html