Update README.md

README.md

@@ -1,58 +1,125 @@
 ---
 license: apache-2.0
 language:
-- zh
 - en
+- zh
 tags:
 - safety
 - moderation
 - multimodal
+- omniguard
 pipeline_tag: text-generation
 ---
 
-# OmniGuard-3B-Fast
+# OmniGuard: Unified Omni-Modal Guardrails with Deliberate Reasoning
+
+OmniGuard is a multimodal safety evaluation model designed to assess content safety across text, images, audio, and video. Built on the Qwen2.5-Omni architecture, it provides structured safety reasoning and policy enforcement.
+
+## Model Information
+
+- **Model Name**: OmniGuard-3B
+- **Base Model**: Qwen2.5-Omni-3B
+- **Model Type**: Multimodal Safety Moderation
+- **Supported Languages**: English, Chinese
+- **Supported Modalities**: Text, Image, Audio, Video
+- **License**: Apache-2.0
+
+## Model Variants
+
+We provide two model sizes:
+
+- **[OmniGuard-3B](https://huggingface.co/anonymousICML/OmniGuard-3B)** - 3B parameters for high-accuracy safety evaluation
+- **[OmniGuard-3B](https://huggingface.co/anonymousICML/OmniGuard-7B)** - 7B parameters for resource-constrained environments
+
+## Features
+
+- **Omni-Modal Safety Assessment**: Evaluate safety across text, images, audio, and video inputs
+- **Structured Reasoning**: Provides deliberate, structured safety analysis beyond binary classification
+
+
+## Installation
+
+### Quick Setup
+
+```bash
+# Install core dependencies
+pip install torch transformers accelerate
 
-OmniGuard 系列安全审核模型。
+# Install Qwen2.5-Omni support
+pip uninstall transformers -y
+pip install git+https://github.com/huggingface/transformers@v4.51.3-Qwen2.5-Omni-preview
 
-## 模型信息
+```
+
+### Requirements
+
+- Python 3.8+
+- PyTorch 2.0+
+- CUDA-compatible GPU (recommended)
+- 8GB+ GPU memory for 3B model
+- Flash Attention 2 (optional, for better performance)
+
+## Usage
 
-- **模型名称**: OmniGuard-3B-Fast
-- **原始模型**: OmniGuard-3B
-- **模型类型**: 多模态安全审核
-- **支持语言**: 中文、英文
 
-## 使用方法
 
 ```python
 from transformers import AutoModelForCausalLM, AutoTokenizer
 
-model_name = "anonymousICML/OmniGuard-3B-Fast"
+model_name = "anonymousICML/OmniGuard-3B"
 model = AutoModelForCausalLM.from_pretrained(
     model_name,
     torch_dtype="auto",
     device_map="auto",
-    attn_implementation="flash_attention_2"  # 推荐使用 flash attention
+    attn_implementation="flash_attention_2"  # Recommended
 )
 tokenizer = AutoTokenizer.from_pretrained(model_name)
 
-# 使用模型
+# Text-only safety evaluation
 messages = [
-    {"role": "user", "content": "你的输入"}
+    {"role": "user", "content": "Please analyze the safety of this content: [Your content here]"}
 ]
-text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+text = tokenizer.apply_chat_template(
+    messages, 
+    tokenize=False, 
+    add_generation_prompt=True
+)
 inputs = tokenizer([text], return_tensors="pt").to(model.device)
 
-outputs = model.generate(**inputs, max_new_tokens=512, do_sample=False)
+outputs = model.generate(
+    **inputs, 
+    max_new_tokens=512, 
+    do_sample=False
+)
 response = tokenizer.decode(outputs[0], skip_special_tokens=True)
 print(response)
 ```
 
-## 性能
 
-- 推荐使用 `flash_attention_2` 以获得最佳性能
-- 支持 GPU 加速推理
-- 建议使用 `torch_dtype="auto"` 自动选择最佳精度
+## Ethical Considerations
+
+- This model is designed for safety evaluation and content moderation
+- Should be used as part of a comprehensive safety strategy
+- May reflect biases present in training data
+- Continuous monitoring and updates recommended for production use
+- Users are responsible for compliance with applicable laws and regulations
 
-## 许可证
+## License
+
+This model is released under the Apache-2.0 license.
+
+## Acknowledgments
+
+- Built on [Qwen2.5-Omni](https://github.com/QwenLM/Qwen2.5-Omni) by Alibaba Cloud
+- Uses [LLaMA Guard 3](https://ai.meta.com/research/publications/llama-guard-llm-based-input-output-safeguard-for-human-ai-conversations/) safety framework
+- Trained on diverse multimodal safety datasets
+
+## Support
+
+For issues, questions, or contributions:
+- Repository: https://github.com/anonymous-2654a/icml-2026-sub
+- HuggingFace: https://huggingface.co/anonymousICML
+
+---
 
-Apache-2.0
+**Note**: This is an anonymous submission for ICML 2026. Full details will be released upon acceptance.