Datasets:

ICLR2024
/

ICLR2024-papers

@@ -11986,12 +11986,8 @@
     "GitHub": [
       "https://github.com/Vchitect/SEINE"
     ],
-    "Space": [
-      "Vchitect/SEINE"
-    ],
-    "Model": [
-      "Vchitect/SEINE"
-    ],
     "Dataset": []
   },
   {
@@ -24012,12 +24008,8 @@
     "GitHub": [
       "https://github.com/guoyww/AnimateDiff"
     ],
-    "Space": [
-      "guoyww/AnimateDiff"
-    ],
-    "Model": [
-      "guoyww/animatediff"
-    ],
     "Dataset": []
   },
   {
@@ -24871,10 +24863,16 @@
     "abstract": "The most advanced text-to-image (T2I) models require significant training costs (e.g., millions of GPU hours), seriously hindering the fundamental innovation for the AIGC community while increasing CO2 emissions. This paper introduces PixArt-$\\alpha$, a Transformer-based T2I diffusion model whose image generation quality is competitive with state-of-the-art image generators (e.g., Imagen, SDXL, and even Midjourney), reaching near-commercial application standards. Additionally, it supports high-resolution image synthesis up to 1024px resolution with low training cost, as shown in Figure 1 and 2. To achieve this goal, three core designs are proposed: (1) Training strategy decomposition: We devise three distinct training steps that separately optimize pixel dependency, text-image alignment, and image aesthetic quality; (2) Efficient T2I Transformer: We incorporate cross-attention modules into Diffusion Transformer (DiT) to inject text conditions and streamline the computation-intensive class-condition branch; (3) High-informative data: We emphasize the significance of concept density in text-image pairs and leverage a large Vision-Language model to auto-label dense pseudo-captions to assist text-image alignment learning. As a result, PixArt-$\\alpha$'s training speed markedly surpasses existing large-scale T2I models, e.g., PixArt-$\\alpha$ only takes 10.8% of Stable Diffusion v1.5's training time (~675 vs. ~6,250 A100 GPU days), saving nearly \\\\$300,000 (\\\\$26,000 vs. \\\\$320,000) and reducing 90% CO2 emissions. Moreover, compared with a larger SOTA model, RAPHAEL, our training cost is merely 1%. Extensive experiments demonstrate that PixArt-$\\alpha$ excels in image quality, artistry, and semantic control. We hope PixArt-$\\alpha$ will provide new insights to the AIGC community and startups to accelerate building their own high-quality yet low-cost generative models from scratch.",
     "type": "Spotlight Poster",
     "OpenReview": "https://openreview.net/forum?id=eAKmQPe3m1",
-    "arxiv_id": "",
-    "GitHub": [],
-    "Space": [],
-    "Model": [],
     "Dataset": []
   },
   {

     "GitHub": [
       "https://github.com/Vchitect/SEINE"
     ],
+    "Space": [],
+    "Model": [],
     "Dataset": []
   },
   {
     "GitHub": [
       "https://github.com/guoyww/AnimateDiff"
     ],
+    "Space": [],
+    "Model": [],
     "Dataset": []
   },
   {
     "abstract": "The most advanced text-to-image (T2I) models require significant training costs (e.g., millions of GPU hours), seriously hindering the fundamental innovation for the AIGC community while increasing CO2 emissions. This paper introduces PixArt-$\\alpha$, a Transformer-based T2I diffusion model whose image generation quality is competitive with state-of-the-art image generators (e.g., Imagen, SDXL, and even Midjourney), reaching near-commercial application standards. Additionally, it supports high-resolution image synthesis up to 1024px resolution with low training cost, as shown in Figure 1 and 2. To achieve this goal, three core designs are proposed: (1) Training strategy decomposition: We devise three distinct training steps that separately optimize pixel dependency, text-image alignment, and image aesthetic quality; (2) Efficient T2I Transformer: We incorporate cross-attention modules into Diffusion Transformer (DiT) to inject text conditions and streamline the computation-intensive class-condition branch; (3) High-informative data: We emphasize the significance of concept density in text-image pairs and leverage a large Vision-Language model to auto-label dense pseudo-captions to assist text-image alignment learning. As a result, PixArt-$\\alpha$'s training speed markedly surpasses existing large-scale T2I models, e.g., PixArt-$\\alpha$ only takes 10.8% of Stable Diffusion v1.5's training time (~675 vs. ~6,250 A100 GPU days), saving nearly \\\\$300,000 (\\\\$26,000 vs. \\\\$320,000) and reducing 90% CO2 emissions. Moreover, compared with a larger SOTA model, RAPHAEL, our training cost is merely 1%. Extensive experiments demonstrate that PixArt-$\\alpha$ excels in image quality, artistry, and semantic control. We hope PixArt-$\\alpha$ will provide new insights to the AIGC community and startups to accelerate building their own high-quality yet low-cost generative models from scratch.",
     "type": "Spotlight Poster",
     "OpenReview": "https://openreview.net/forum?id=eAKmQPe3m1",
+    "arxiv_id": "2310.00426",
+    "GitHub": [
+      "https://github.com/PixArt-alpha/PixArt-alpha"
+    ],
+    "Space": [
+      "PixArt-alpha/PixArt-alpha"
+    ],
+    "Model": [
+      "PixArt-alpha/PixArt-XL-2-1024-MS"
+    ],
     "Dataset": []
   },
   {