Update app.py

app.py

@@ -1,5 +1,4 @@
 # This is a project of Chakra Lab LLC. All rights reserved.
-import spaces
 
 import gradio as gr
 import os
@@ -22,19 +21,14 @@ bnb_config = BitsAndBytesConfig(
     #bnb_4bit_use_double_quant=True
 )
 
-@spaces.GPU
-def load_model():
-    model = AutoModelForCausalLM.from_pretrained(base_model_name,
-                                                 token=os.environ['HF_TOKEN'],
-                                                 quantization_config=bnb_config,
-                                                 device_map="auto")
-    model = PeftModel.from_pretrained(model, adapter_model_name, token=os.environ['HF_TOKEN'])
-    model.merge_and_unload()
-    
-    model = model.to(device)
-    return model
+model = AutoModelForCausalLM.from_pretrained(base_model_name,
+                                             token=os.environ['HF_TOKEN'],
+                                             quantization_config=bnb_config,
+                                             device_map="auto")
+model = PeftModel.from_pretrained(model, adapter_model_name, token=os.environ['HF_TOKEN'])
+model.merge_and_unload()
 
-model = load_model()
+model = model.to(device)
 
 tokenizer = AutoTokenizer.from_pretrained(base_model_name)
 
@@ -97,7 +91,6 @@ This policy is designed to determine whether or not content is hate speech.
 DEFAULT_CONTENT = "LLMs steal our jobs."
 
 # Function to make predictions
-@spaces.GPU
 def predict(content, policy):
     input_text = PROMPT.format(policy=policy, content=content)
     input_ids = tokenizer.encode(input_text, return_tensors="pt")