SamuelBang
/

AesCoder-4B

@@ -1,11 +1,16 @@
 ---
 base_model:
 - Qwen/Qwen3-4B-Instruct-2507
 ---
 # Code Aesthetics with Agentic Reward Feedback
 <div align="center">
 <a href='https://bangx7.github.io/code-aesthetics/'><img src='https://img.shields.io/badge/Project-Page-Green'></a>
   <br>
   <a href="https://arxiv.org/abs/2510.23272"><b>Paper Link</b>👁️</a>
 </div>
@@ -41,6 +46,19 @@ base_model:
   </p>
 </div>
 **Note: This is the version of AesCoder-4B model for only webpage design.**
 ## Quickstart
@@ -295,9 +313,6 @@ Return ONLY the full, standalone HTML code (starting with <!DOCTYPE html>) and n
 If the user specifies a particular style (e.g., glassmorphism, brutalism, Material Design), follow their style instructions instead of the default design preferences.
 ```
 ## &#x1F4DA; Citation
 If you find this codebase useful for your research, please use the following entry.
 ```BibTeX

 ---
 base_model:
 - Qwen/Qwen3-4B-Instruct-2507
+pipeline_tag: text-generation
+library_name: transformers
+license: apache-2.0
 ---
 # Code Aesthetics with Agentic Reward Feedback
 <div align="center">
 <a href='https://bangx7.github.io/code-aesthetics/'><img src='https://img.shields.io/badge/Project-Page-Green'></a>
+  <a href="https://huggingface.co/SamuelBang/AesCoder-4B"><img alt="Hugging Face"
+    src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Model-ffc107?color=ffc107&logoColor=white"/></a>
   <br>
   <a href="https://arxiv.org/abs/2510.23272"><b>Paper Link</b>👁️</a>
 </div>
   </p>
 </div>
+For the codebase, refer to: https://github.com/bangx7/code_aesthetics
+## 🎉 News
+- __[2025.10.27]__: Release the [Project Page](https://bangx7.github.io/code-aesthetics/) and the [Arxiv](https://arxiv.org/abs/2510.23272) version.
+## 📷 Abstract
+Large Language Models (LLMs) have become valuable assistants for developers in code-related tasks. While LLMs excel at traditional programming tasks such as code generation and bug fixing, they struggle with visually-oriented coding tasks, often producing suboptimal aesthetics. In this paper, we introduce a new pipeline to enhance the aesthetic quality of LLM-generated code. We first construct AesCode-358K, a large-scale instruction-tuning dataset focused on code aesthetics. Next, we propose agentic reward feedback, a multi-agent system that evaluates executability, static aesthetics, and interactive aesthetics. Building on this, we develop GRPO-AR, which integrates these signals into the GRPO algorithm for joint optimization of functionality and code aesthetics. Finally, we develop OpenDesign, a benchmark for assessing code aesthetics. Experimental results show that combining supervised fine-tuning on AesCode-358K with reinforcement learning using agentic reward feedback significantly improves performance on OpenDesign and also enhances results on existing benchmarks such as PandasPlotBench. Notably, our AesCoder-4B surpasses GPT-4o and GPT-4.1, and achieves performance comparable to large open-source models with 480B-685B parameters, underscoring the effectiveness of our approach.
+## To-do List
+- [x] Release paper and project page
+- [ ] Release our AesCoder model
+- [ ] Release code
 **Note: This is the version of AesCoder-4B model for only webpage design.**
 ## Quickstart
 If the user specifies a particular style (e.g., glassmorphism, brutalism, Material Design), follow their style instructions instead of the default design preferences.
 ```
 ## &#x1F4DA; Citation
 If you find this codebase useful for your research, please use the following entry.
 ```BibTeX