Update README.md

README.md

@@ -98,30 +98,19 @@ model = xlora.from_pretrained(
     "cuda",
 )
 ```
-## Loading pre-trained X-LoRA model
+## Loading pre-trained X-LoRA model directly from Hugging Face Hub
 
 ```python
 import torch
-from xlora.xlora_utils import load_model  # type: ignore
+from xlora.xlora_utils import load_model  
 
-XLoRA_model_name = "lamm-mit/x-lora/X-LoRA"
+XLoRa_model_name = 'lamm-mit/x-lora'
 
-model, tokenizer = load_model(
-    model_name="HuggingFaceH4/zephyr-7b-beta",
-    device="cuda:0",
-    dtype=torch.bfloat16,
-    fine_tune_model_name=XLoRA_model_name,
-    adapters={
-        "adapter_1": "lamm-mit/x-lora/X-LoRA_adapters/1/",
-        "adapter_2": "lamm-mit/x-lora/X-LoRA_adapters/2/",
-        "adapter_3": "lamm-mit/x-lora/X-LoRA_adapters/3/",
-        "adapter_4": "lamm-mit/x-lora/X-LoRA_adapters/4/",
-        "adapter_5": "lamm-mit/x-lora/X-LoRA_adapters/5/",
-        "adapter_6": "lamm-mit/x-lora/X-LoRA_adapters/6/",
-        "adapter_7": "lamm-mit/x-lora/X-LoRA_adapters/7/",
-        "adapter_8": "lamm-mit/x-lora/X-LoRA_adapters/8/",
-        "adapter_9": "lamm-mit/x-lora/X-LoRA_adapters/9/",
-    },
+model,tokenizer=load_model(model_name = XLoRa_model_name, 
+                           device='cuda:0',
+                           use_flash_attention_2=True, 
+                           dtype=torch.bfloat16,
+                            )
 )
 ```
 Inference: