Upload folder using huggingface_hub

README.md

@@ -9,36 +9,30 @@ language:
 base_model:
 - Qwen/Qwen2.5-32B
 ---
-<div align="center">
-
-# Baichuan-M2-32B
+# Baichuan-M2-32B-GPTQ-Int4
 
 [![License](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://opensource.org/licenses/Apache-2.0)
 [![Hugging Face](https://img.shields.io/badge/🤗%20Hugging%20Face-Model-yellow)](https://huggingface.co/baichuan-inc/Baichuan-M2-32B)
 
-</div>
-
-## 🌟 模型简介
-
-Baichuan-M2-32B 是百川智能推出的医疗增强推理模型，这是百川开源发布的第二个医疗增强模型，专为真实世界的医疗推理任务设计。该模型基于 Qwen2.5-32B 基座，通过创新的大型验证器系统（Large Verifier System）从真实世界的医疗问题出发，进行医疗领域后训练对齐，在保持模型通用能力的同时，实现了医疗效果的突破性提升。
+## 🌟 Model Overview
 
-**模型特点：**
+Baichuan-M2-32B is BaiChuan AI's medical-enhanced reasoning model, the second medical model released by BaiChuan. Designed for real-world medical reasoning tasks, this model builds upon Qwen2.5-32B with an innovative Large Verifier System. Through domain-specific fine-tuning on real-world medical questions, it achieves breakthrough medical performance while maintaining strong general capabilities.
 
-Baichuan-M2 采用了三个核心技术创新：首先通过**大型验证器系统**，结合医疗场景特点设计了全面的医疗验证体系，包含患者模拟器和多维度验证机制；其次通过**医疗领域适应性增强**的中期训练（Mid-Training），在保持通用能力的同时实现轻量高效的医疗领域适应；最后采用**多阶段强化学习**策略，将复杂的 RL 任务分解为层次化的训练阶段，逐步提升模型的医学常识、推理和患者交互能力。
+**Model Features:**
 
-**核心亮点：**
-- 🏆 **全球最强医疗开源模型**：在 HealthBench 评测集上超越所有开源模型及众多前沿闭源模型，是最接近 GPT-5 医疗能力的开源大模型
-- 🧠 **医生思维对齐**：基于真实病例数据和患者模拟器训练，具备临床诊断思维和鲁棒的医患交互能力
-- ⚡ **高效部署与推理**：支持 4bit 量化在 RTX4090 单卡部署，MTP 版本单用户场景下 token 吞吐提升 58.5%
+Baichuan-M2 incorporates three core technical innovations: First, through the **Large Verifier System**, it combines medical scenario characteristics to design a comprehensive medical verification framework, including patient simulators and multi-dimensional verification mechanisms; second, through **medical domain adaptation enhancement** via Mid-Training, it achieves lightweight and efficient medical domain adaptation while preserving general capabilities; finally, it employs a **multi-stage reinforcement learning** strategy, decomposing complex RL tasks into hierarchical training stages to progressively enhance the model's medical knowledge, reasoning, and patient interaction capabilities.
 
+**Core Highlights:**
+- 🏆 **World's Leading Open-Source Medical Model**: Outperforms all open-source models and many proprietary models on HealthBench, achieving medical capabilities closest to GPT-5
+- 🧠 **Doctor-Thinking Alignment**: Trained on real clinical cases and patient simulators, with clinical diagnostic thinking and robust patient interaction capabilities
+- ⚡ **Efficient Deployment**: Supports 4-bit quantization for single-RTX4090 deployment, with 58.5% higher token throughput in MTP version for single-user scenarios
 
+## 📊 Performance Metrics
 
-## 📊 性能表现
+### HealthBench Scores
 
-### HealthBench指标
-
-| 模型名称 | HealthBench | HealthBench-Hard | HealthBench-Consensus |
-|----------|-------------|------------------|-----------------------|
+| Model Name | HealthBench | HealthBench-Hard | HealthBench-Consensus |
+|------------|-------------|------------------|-----------------------|
 | Baichuan-M2 | 60.1 | 34.7 | 91.5 |
 | gpt-oss-120b | 57.6 | 30 | 90 |
 | Qwen3-235B-A22B-Thinking-2507 | 55.2 | 25.9 | 90.6 |
@@ -47,80 +41,83 @@ Baichuan-M2 采用了三个核心技术创新：首先通过**大型验证器系
 | Kimi-K2 | 43 | 10.7 | 90.9 |
 | gpt-oss-20b | 42.5 | 10.8 | 82.6 |
 
-### 通用指标
+### General Performance
 
-| 评测集 | Baichuan-M2-32B | Qwen3-32B |
-|--------|-----------------|-----------|
+| Benchmark | Baichuan-M2-32B | Qwen3-32B |
+|-----------|-----------------|-----------|
 | AIME24 | 83.4 | 81.4 |
 | AIME25 | 72.9 | 72.9 |
 | Arena-Hard-v2.0 | 45.8 | 44.5 |
 | CFBench | 77.6 | 75.7 |
 | WritingBench | 8.56 | 7.90 |
 
-*备注：AIME 的 max_tokens 设为 64k，其他评测集设为 32k，温度统一为 0.6。*
-
-
-## 🛠️ 技术特色
-
-### 大型验证器系统
-- **患者模拟器**：基于真实病例构建的虚拟患者系统
-- **多维度验证**：医学准确性、回答完整性、追问感知等 8 个维度
-- **动态评分**：实时生成评分标准，适应复杂临床环境
+*Note: AIME uses max_tokens=64k, others use 32k; temperature=0.6 for all tests.*
 
-### 医疗领域适应
-- **Mid-Training**：医疗知识注入的同时保持通用能力
-- **强化学习**：多阶段 RL 策略优化
-- **通专兼顾**：2:2:1 配比的医疗、通用、数学数据
+## 🔧 Technical Features
 
-## 🔧 快速开始
+### Large Verifier System
+- **Patient Simulator**: Virtual patient system based on real clinical cases
+- **Multi-Dimensional Verification**: 8 dimensions including medical accuracy, response completeness, and follow-up awareness
+- **Dynamic Scoring**: Real-time generation of adaptive evaluation criteria for complex clinical scenarios
+### Medical Domain Adaptation
+- **Mid-Training**: Medical knowledge injection while preserving general capabilities
+- **Reinforcement Learning**: Multi-stage RL strategy optimization
+- **General-Specialized Balance**: Carefully balanced medical, general, and mathematical composite training data
 
-### 安装使用
+## ⚙️ Quick Start
 
-```bash
-# 安装依赖
-pip install transformers torch vllm sglang
+For deployment, you can use `sglang>=0.4.6.post1` or `vllm>=0.9.0` or to create an OpenAI-compatible API endpoint:
+- SGLang:
+    ```shell
+    python -m sglang.launch_server --model-path baichuan-inc/Baichuan-M2-32B-GPTQ-Int4 --reasoning-parser qwen3
+    ```
+- vLLM:
+    ```shell
+    vllm serve baichuan-inc/Baichuan-M2-32B-GPTQ-Int4  --reasoning-parser qwen3
+    ```
 
-# Transformers 使用
-from transformers import AutoTokenizer, AutoModelForCausalLM
-model = AutoModelForCausalLM.from_pretrained("baichuan-inc/Baichuan-M2-32B", trust_remote_code=True)
+## MTP inference with SGLang
 
-# vLLM 使用（推荐）
-from vllm import LLM
-llm = LLM(model="baichuan-inc/Baichuan-M2-32B", trust_remote_code=True)
-
-# SGLang 使用
-python -m sglang.launch_server --model-path baichuan-inc/Baichuan-M2-32B
+1. Replace the qwen2.py file in the sglang installation directory with draft/qwen2.py.
+2. Launch sglang:
+```
+python3 -m sglang.launch_server \
+--model Baichuan-M2-32B-GPTQ-Int4 \
+--speculative-algorithm EAGLE3 \
+--speculative-draft-model-path Baichuan-M2-32B-GPTQ-Int4/draft \
+--speculative-num-steps 6 \
+--speculative-eagle-topk 10 \
+--speculative-num-draft-tokens 32 \
+--mem-fraction 0.9 \
+--cuda-graph-max-bs 2 \
+--reasoning-parser qwen3 \
+--dtype bfloat16
 ```
 
-## ⚠️ 使用须知
-
-1. **医疗免责声明**：本模型仅供研究和参考，不能替代专业医疗诊断和治疗建议
-2. **适用场景**：医学教育、健康咨询、临床辅助决策等
-3. **安全使用**：建议在专业医疗人员指导下使用
-
-## 📄 许可证
-
-本项目采用 [Apache License 2.0](LICENSE) 开源协议，欢迎研究和商业使用。
-
-## 🤝 致谢
-
-- 基础模型：Qwen2.5-32B
-- 训练框架：VERL
-- 推理引擎：vLLM、SGLang
-- 量化方法：AutoRound、GPTQ
+## ⚠️ Usage Notices
+1. **Medical Disclaimer**: For research and reference only; cannot replace professional medical diagnosis or treatment
+2. **Intended Use Cases**: Medical education, health consultation, clinical decision support
+3. **Safe Use**: Recommended under guidance of medical professionals
 
-感谢开源社区的贡献，我们将持续回馈社区，推动医疗 AI 技术发展。
+## 📄 License
+Licensed under the [Apache License 2.0](LICENSE). Research and commercial use permitted.
 
-## 📞 联系我们
+## 🤝 Acknowledgements
+- Base Model: Qwen2.5-32B
+- Training Framework: verl
+- Inference Engines: vLLM, SGLang
+- Quantization: AutoRound, GPTQ
 
-- 更多资源：[百川智能官网](https://www.baichuan-ai.com)
+Thank you to the open-source community. We commit to continuous contribution and advancement of healthcare AI.
 
-- 技术交流：[GitHub](https://github.com/baichuan-inc)
+## 📞 Contact Us
+- Resources: [BaiChuan AI Website](https://www.baichuan-ai.com)
+- Technical Support: [GitHub](https://github.com/baichuan-inc)
 
 ---
-
 <div align="center">
 
-**让AI助力医疗，让健康触手可及**
+**Empowering Healthcare with AI, Making Health Accessible to All**
 
 </div>
+