Update README.md

README.md

@@ -9,7 +9,7 @@ library_name: peft
 ---
 
 This is the B2NER model's LoRA adapter based on [InternLM2-20B](https://huggingface.co/internlm/internlm2-20b).  
-**See [github repo](https://github.com/UmeanNever/B2NER) for quick demo usage and more information about this work.**
+**See our [GitHub Repo](https://github.com/UmeanNever/B2NER) for model usage and more information about this work.**
 
 ## B2NER 
 
@@ -21,6 +21,14 @@ Our B2NER models, trained on B2NERD, outperform GPT-4 by 6.8-12.0 F1 points and
  - 📀 Data: See [B2NERD](https://huggingface.co/datasets/Umean/B2NERD).
  - 💾 Model (LoRA Adapters): Current repo saves the B2NER model LoRA adapter based on InternLM2-20B. See [7B model](https://huggingface.co/Umean/B2NER-Internlm2.5-7B-LoRA) for a 7B adapter.
 
+**Feature Highlights:**
+ - Curated dataset (B2NERD) refined from the largest bilingual NER dataset collection to date for training Open NER models.  
+ - Achieves SoTA OOD NER performance across multiple benchmarks with light-weight LoRA adapters (<=50MB). 
+ - Uses simple natural language format prompt, achieving 4X faster inference speed than previous SoTA which use complex prompts. 
+ - Easy integration with other IE tasks by adopting UIE-style instructions. 
+ - Provides a universal entity taxonomy that guides the definition and label naming of new entities.
+ - We have open-sourced our data, code, and models, and provided easy-to-follow usage instructions. 
+
 ## Sample Usage - Quick Demo
 Here we show how to use our provided lora adapter to do quick demo with customized input. You can also refer to github repo's `src/demo.ipynb` to see our examples and reuse for your own demo.
  - Prepare/download our LoRA checkpoint and corresponding backbone model.