Update README.md

README.md

@@ -1,3 +1,156 @@
 ---
 license: apache-2.0
+base_model: Qwen/Qwen3-8B
+library_name: transformers
+pipeline_tag: text-generation
+language:
+- en
+- zh
+tags:
+- finance
+- reinforcement-learning
+- reasoning
+- qwen3
+- stock-selection
+- quantitative-finance
 ---
+
+# Alpha-R1
+
+<p align="center">
+  <img src="https://huggingface.co/front/assets/huggingface_logo.svg" width="120">
+</p>
+
+**Alpha-R1** is a reasoning-enhanced Large Language Model for quantitative stock selection, trained with Reinforcement Learning (RL) on top of **Qwen3-8B**.
+
+It is the official implementation accompanying the paper:
+
+> [Alpha-R1: Alpha Screening with LLM Reasoning via Reinforcement Learning](https://arxiv.org/abs/2512.23515)
+
+---
+
+# Overview
+
+Alpha-R1 is designed for **Alpha Screening** rather than general-purpose conversation.
+
+Unlike conventional LLMs, Alpha-R1 learns to reason over
+
+- Financial factor descriptions
+- Historical price trends
+- Market news
+- Portfolio constraints
+
+to produce interpretable alpha-selection decisions.
+
+The model is optimized using reinforcement learning with trading performance as the optimization objective, enabling stronger reasoning ability for quantitative investment tasks.
+
+---
+
+# Highlights
+
+- 🧠 Reinforcement-learning aligned financial reasoning
+- 📈 Multi-modal market understanding (news + quantitative factors + price)
+- 📊 Strong generalization across different asset pools
+- 💰 Optimized for alpha generation instead of language modeling
+
+According to the paper, Alpha-R1 achieves:
+
+| Dataset | Annual Return | Sharpe | Max Drawdown |
+|----------|--------------|---------|--------------|
+| CSI300 | 27.59% | 1.62 | 6.76% |
+| CSI1000 | 78.18% | 4.03 | 9.25% |
+
+---
+
+# Model Details
+
+| Item | Value |
+|------|------|
+| Base Model | Qwen3-8B |
+| Model Type | Causal Language Model |
+| Training | Reinforcement Learning Fine-tuning |
+| Domain | Quantitative Finance |
+| Language | English |
+| Intended Task | Stock Selection & Financial Reasoning |
+
+---
+
+# Training
+
+Alpha-R1 is initialized from **Qwen3-8B** and further optimized using reinforcement learning.
+
+The training objective encourages the model to generate reasoning trajectories that maximize downstream portfolio performance instead of only predicting next tokens.
+
+The model is trained using:
+
+- Financial factor descriptions
+- Historical price information
+- Market news
+- Trading rewards derived from portfolio returns
+
+More details can be found in the accompanying paper.
+
+---
+
+## Training Data
+
+Alpha-R1 was trained using a proprietary financial reasoning dataset constructed from market information, factor semantic descriptions, and financial news. Due to data licensing and copyright restrictions, the training dataset is not publicly released.
+
+---
+
+# Intended Use
+
+Alpha-R1 is intended for research on
+
+- Financial reasoning
+- Quantitative investment
+- Alpha mining
+- LLM-based decision making
+- Reinforcement learning for finance
+
+Example prompts include:
+
+- Rank candidate stocks.
+- Explain factor exposure.
+- Analyze market news.
+- Compare investment opportunities.
+- Generate reasoning for stock selection.
+
+---
+
+# Limitations
+
+- This model is **not** a financial advisor.
+- Outputs should **not** be regarded as investment advice.
+- Performance reported in the paper is obtained under a specific backtesting protocol and does not guarantee future returns.
+- Users should perform their own validation before any real-world deployment.
+
+---
+
+# Citation
+
+If you find Alpha-R1 useful, please cite:
+
+```bibtex
+@article{alphar1,
+  title={Alpha-R1: Alpha Screening with LLM Reasoning via Reinforcement Learning},
+  author={Anonymous},
+  year={2025}
+}
+```
+
+---
+
+# License
+
+This model is released under the Apache-2.0 License.
+
+Please also comply with the license of the base model (**Qwen3-8B**) when using this model.
+
+---
+
+# Acknowledgements
+
+Alpha-R1 is built upon the excellent **Qwen3-8B** model developed by Alibaba.
+
+We thank the open-source community for making this work possible.