Update README.md

README.md

@@ -15,23 +15,18 @@ A BPE tokenizer that extends the LLaMA tokenizer vocabulary with all Unicode CJK
 - **Character-Level**  
   - Each CJK character is treated as its own token  
   - Other scripts (Latin, punctuation, etc.) follow the original LLaMA subword scheme
-- **Hugging Face Integration**  
-  - Fully compatible with the `transformers` library  
-  - Supports `from_pretrained`, `save_pretrained`, and standard tokenizer methods
 
 ## Installation
 
-
+```python
 from transformers import AutoTokenizer
 
-# Load the tokenizer by repository name or local path
 tokenizer = AutoTokenizer.from_pretrained("AgentBull/CJK-Tokenizer")
 
-# Tokenize some CJK text
 text = "天地玄黃， 宇宙洪荒。 日月盈昃， 辰宿列張。"
 tokens = tokenizer(text)
 print(tokens.tokens)
-
+```
 
 
 Tokenizer Details