Changes in app.py

app.py

@@ -13,17 +13,18 @@ model = "meta-llama/Llama-2-13b-chat-hf"
 #Not Working
 #tokenizer = AutoTokenizer.from_pretrained(model)
 
+def load_model(model):
+    pipeline = transformers.pipeline(
+        "text-generation",
+        model=model,
+        torch_dtype=torch.float32,
+        device_map="auto",
+        do_sample=True,
+        token=HF_TOKEN,
+    )
+    return pipeline
 
-pipeline = transformers.pipeline(
-    "text-generation",
-    model=model,
-    torch_dtype=torch.float32,
-    device_map="auto",
-    do_sample=True,
-    token=HF_TOKEN,
-)
-
-def get_llama_response(prompt):
+def get_llama_response(pipeline,prompt):
     
     sequences = pipeline(
         prompt,
@@ -33,8 +34,10 @@ def get_llama_response(prompt):
         max_length=256,
     )
     print(sequences[0]['generated_text'])
+    
 
+pipeline = AutoTokenizer.from_pretrained(model)
 #prompt="Can you help me to write rest api endpoints in python ?"
-response = get_llama_response(prompt)
+response = get_llama_response(pipeline,prompt)
 
 st.write('Answer: ',response)