Add model card and metadata

README.md

@@ -0,0 +1,44 @@
+---
+library_name: transformers
+pipeline_tag: text-generation
+base_model: Qwen/Qwen3-8B
+---
+
+# EvoDS: Self-Evolving Autonomous Data Science Agent
+
+EvoDS is a self-evolving autonomous data science agent designed to address limitations in LLM-based data science systems through skill learning and adaptive context management. It was introduced in the paper [EvoDS: Self-Evolving Autonomous Data Science Agent with Skill Learning and Context Management](https://huggingface.co/papers/2606.03841).
+
+- **Repository:** [https://github.com/usail-hkust/EvoDS](https://github.com/usail-hkust/EvoDS)
+- **Paper:** [https://huggingface.co/papers/2606.03841](https://huggingface.co/papers/2606.03841)
+
+## Overview
+
+EvoDS introduces two key strategies to overcome the limitations of static action sets and long-horizon context management in data science tasks:
+
+1.  **Autonomous Skill Acquisition (ASA):** This mechanism enables the agent to autonomously synthesize, validate, and internalize reusable tool-usage skills from its experience.
+2.  **Adaptive Context Compression (ACC):** This strategy treats context management as a learned control problem, dynamically compressing interaction history to maintain reasoning stability within limited context budgets.
+
+The model is optimized using agentic reinforcement learning to jointly improve task completion quality, skill acquisition behavior, and context regulation.
+
+## Performance
+
+EvoDS outperforms state-of-the-art open-source data science agents by an average of 28.9% across four diverse benchmarks:
+- DABench
+- DA-Code
+- ScienceAgentBench
+- MLE-Dojo
+
+## Usage
+
+For detailed instructions on how to deploy and evaluate the agent using vLLM or VERL, please refer to the [official GitHub repository](https://github.com/usail-hkust/EvoDS).
+
+## Citation
+
+```bibtex
+@inproceedings{yang2026evods,
+  title={EvoDS: Self-Evolving Autonomous Data Science Agent with Skill Learning and Context Management},
+  author={Yang, Zherui and Liu, Fan and Ning, Yansong and Liu, Hao},
+  booktitle={Proceedings of the 32nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD)},
+  year={2026}
+}
+```