Keep model card icon links inline on Hugging Face

README.md

@@ -13,19 +13,17 @@ tags:
   - adversarial-gradient-forcing
 ---
 
-<p align="center">
-  <img src="assets/title.svg" width="960" alt="Adversarial Gradient Forcing — Robust Few-Step Autoregressive Video Generation">
+<p align="center" style="margin: 0 0 12px;">
+  <img src="assets/title.svg" width="960" style="display: block; margin: 0 auto;" alt="Adversarial Gradient Forcing — Robust Few-Step Autoregressive Video Generation">
 </p>
-<p align="center">
+<p align="center" style="margin: 8px 0 14px;">
   <sub>Zhejiang University &nbsp;·&nbsp; Alibaba Group &nbsp;·&nbsp; Nanjing University</sub>
 </p>
-<p align="center">
-  <a href="https://github.com/wuli-art/Adversarial-Gradient-Forcing"><img src="assets/github.svg" height="38" alt="GitHub — source code"></a>
-  &nbsp;
-  <a href="https://shark-console.oss-cn-shanghai.aliyuncs.com/shuju/project/kv_reasoning/runs/AGF/reports/project-page/20261010-explainer-ea1f9094/index.html"><img src="assets/project.svg" height="38" alt="Project page — videos and results"></a>
-  &nbsp;
-  <img src="assets/arxiv.svg" height="38" alt="arXiv — coming soon">
-</p>
+<div align="center" style="line-height: 1.4; margin: 12px 0 22px;">
+  <a href="https://github.com/wuli-art/Adversarial-Gradient-Forcing"><img src="assets/github.svg" height="38" style="display: inline-block; vertical-align: middle; margin: 4px 3px;" alt="GitHub — source code"></a>
+  <a href="https://shark-console.oss-cn-shanghai.aliyuncs.com/shuju/project/kv_reasoning/runs/AGF/reports/project-page/20261010-explainer-ea1f9094/index.html"><img src="assets/project.svg" height="38" style="display: inline-block; vertical-align: middle; margin: 4px 3px;" alt="Project page — videos and results"></a>
+  <img src="assets/arxiv.svg" height="38" style="display: inline-block; vertical-align: middle; margin: 4px 3px;" alt="arXiv — coming soon">
+</div>
 
 **AGF** improves long-video generation by training models to withstand **adversarial perturbations in their generated history**. Trained on five-second rollouts, it supports generation up to four minutes with the same four-step sampler.