update gradio

UltraEdit

@@ -0,0 +1 @@
+Subproject commit 1f07d3cf2026ef372d3117f0bede5240f5e0e850

app.py

@@ -65,16 +65,44 @@ outputs = gr.Image(label="Generated Image")
 
 
 # Custom HTML content
-article_html = """
-<h2>Welcome to the Image Generation Interface</h2>
-<p>This interface allows you to generate images based on a given mask and prompt. Use the sliders to adjust the inference steps and guidance scales, and provide a seed for reproducibility.</p>
-"""
+# article_html = """
+# <div style="text-align: center; max-width: 1200px; margin: 20px auto;">
+#   <h1 style="font-weight: 900; font-size: 3rem; margin-bottom: 0.5rem">🖼️ Stable Diffusion 3 Image Editor</h1>
+#   <h2 style="font-weight: 450; font-size: 1rem; margin: 0rem">
+#     This interface allows you to perform image editing using the Stable Diffusion 3 model trained with the UltraEdit dataset.
+#   </h2>
+#   <h2 style="font-weight: 450; font-size: 1rem; margin: 0.7rem auto; max-width: 1000px">
+#     Supports both free-form (without mask) and region-based (with mask) image editing. Use the sliders to adjust the inference steps and guidance scales, and provide a seed for reproducibility.
+#   </h2>
+#   <h2 style="font-weight: 450; font-size: 1rem; margin: 1rem auto; max-width: 1000px">
+#     <b>UltraEdit: Instruction-based Fine-Grained Image Editing at Scale</b>
+#   </h2>
+#   <div style="text-align: left; max-width: 1000px; margin: 0 auto;">
+#     <p>
+#       Haozhe Zhao<sup>1*</sup>, Xiaojian Ma<sup>2*</sup>, Liang Chen<sup>1</sup>, Shuzheng Si<sup>1</sup>, Rujie Wu<sup>1</sup>, Kaikai An<sup>1</sup>, Peiyu Yu<sup>3</sup>, Minjia Zhang<sup>4</sup>, Qing Li<sup>2</sup>, Baobao Chang<sup>2</sup>
+#       <br>
+#       <sup>1</sup>Peking University, <sup>2</sup>BIGAI, <sup>3</sup>UCLA, <sup>4</sup>UIUC
+#     </p>
+#     <p>
+#       This paper presents ULTRAEDIT, a large-scale (~4M editing samples), automatically generated dataset for instruction-based image editing. Our key idea is to address the drawbacks in existing image editing datasets like InstructPix2Pix and MagicBrush, and provide a systematic approach to producing massive and high-quality image editing samples. ULTRAEDIT offers several distinct advantages:
+#     </p>
+#     <ul>
+#       <li>It features a broader range of editing instructions by leveraging the creativity of large language models (LLMs) alongside in-context editing examples from human raters.</li>
+#       <li>Its data sources are based on real images, including photographs and artworks, which provide greater diversity and reduced bias compared to datasets solely generated by text-to-image models.</li>
+#       <li>It also supports region-based editing, enhanced by high-quality, automatically produced region annotations.</li>
+#     </ul>
+#     <p>
+#       Our experiments show that canonical diffusion-based editing baselines trained on ULTRAEDIT set new records on MagicBrush and Emu-Edit benchmarks. Our analysis further confirms the crucial role of real image anchors and region-based editing data. The dataset, code, and models will be made public.
+#     </p>
+#   </div>
+# </div>
+# """
 
 demo = gr.Interface(
     fn=generate,
     inputs=inputs,
     outputs=outputs,
-    article=article_html  # Add article parameter
+    # title=article_html  # Add article parameter
 )
 
 demo.queue().launch()