Update app.py

app.py

@@ -12,7 +12,7 @@ model_path = snapshot_download(
 )
 
 tokenizer = LlamaTokenizer.from_pretrained(model_path)
-model = LlamaForCausalLM.from_pretrained(model_path)
+model = LlamaForCausalLM.from_pretrained(model_path,device_map='auto')
 
 DESCRIPTION = '''
 # SimpleBerry/LLaMA-O1-Supervised-1129 | Duplicate the space and set it to private for faster & personal inference for free.
@@ -35,6 +35,7 @@ def llama_o1_template(data):
     text = template.format(content=data)
     return text
 
+@spaces.GPU
 def generate_text(message, history, max_tokens=512, temperature=0.9, top_p=0.95):
     input_text = llama_o1_template(message)
     inputs = tokenizer(input_text, return_tensors="pt")