adapter description updated

README.md

@@ -44,7 +44,7 @@ Install the required dependencies:
 pip install transformers torch numpy pandas anarci
 ```
 
-## 🚀 Loading Models from Hugging Face Hub
+## 🚀 Loading Model from Hugging Face Hub
 
 ### Method 1: Load Model and Tokenizer, then Import Adapter
 ```python
@@ -87,10 +87,12 @@ spec.loader.exec_module(adapter_module)
 ablang = adapter_module.AbLang2PairedHuggingFaceAdapter(model=model, tokenizer=tokenizer)
 ```
 
-**Note**: Models automatically use GPU when available, otherwise fall back to CPU.
+**Note**: Model automatically use GPU when available, otherwise fall back to CPU.
 
 ## ⚙️ Available Utilities
 
+This wrapper translates between HuggingFace's model format and AbLang2's expected input/output structure, making it easy to use AbLang2's powerful antibody analysis tools with model loaded from HuggingFace.
+
 - **seqcoding**: Sequence-level representations (averaged across residues)
 - **rescoding**: Residue-level representations (per-residue embeddings)
 - **likelihood**: Raw logits for amino acid prediction at each position
@@ -99,6 +101,14 @@ ablang = adapter_module.AbLang2PairedHuggingFaceAdapter(model=model, tokenizer=t
 - **confidence**: Fast uncertainty scoring (single forward pass, no masking)
 - **restore**: Restore masked residues (*) with predicted amino acids
 
+All these utilities work seamlessly with the HuggingFace-loaded model, maintaining the same API as the original AbLang2 implementation.
+
+The `AbLang2PairedHuggingFaceAdapter` class is a wrapper that lets you use AbLang2 model utilities after loading the model from HuggingFace. This class enables you to:
+
+- **Access all AbLang2 utilities** (seqcoding, rescoding, likelihood, probability, etc.) with the same interface as the original implementation
+- **Work with antibody sequences** (heavy and light chains) seamlessly
+- **Maintain compatibility** with the original AbLang2 API while leveraging HuggingFace's model loading and caching capabilities
+
 ## 💡 Examples
 
 ### 🔗 AbLang2 (Paired Sequences) - Restore Example