Update README.md

README.md

@@ -18,6 +18,20 @@ The models are prepared by [Visual Informatics Group @ University of Texas at Au
 
 License: [MIT License](https://opensource.org/license/mit/)
 
+How to use
+```python
+base_model = 'llama-2-7b'
+comp_degree = 0.1
+comp_method = 'sparsegpt_unstructured'
+model_path = f'vita-group/comp-{arch}_{comp_method}_s{comp_degree}'
+model = AutoModelForCausalLM.from_pretrained(
+        model_path, 
+        torch_dtype=torch.float16, 
+        low_cpu_mem_usage=True, 
+        device_map="auto"
+    )
+```
+
 
 |    | Base Model   | Model Size   | Compression Method     | Compression Degree                                                                    |
 |---:|:-------------|:-------------|:-----------------------|:--------------------------------------------------------------------------------------|