Update app.py

app.py

@@ -11,36 +11,15 @@ def greet_json():
 @app.get("/hello/{msg}")
 def say_hello(msg: str):
     print("model")
-    torch.random.manual_seed(0) 
-    model = AutoModelForCausalLM.from_pretrained( 
-        "microsoft/Phi-3-mini-4k-instruct",  
-        device_map="auto",  
-        torch_dtype="auto",  
-        trust_remote_code=True,  
-    ) 
+    tokenizer = AutoTokenizer.from_pretrained("google/gemma-2b-it")
+    model = AutoModelForCausalLM.from_pretrained(
+        "google/gemma-2b-it",
+        device_map="auto",
+        torch_dtype=torch.bfloat16
+    )
     print("token & msg")
-    tokenizer = AutoTokenizer.from_pretrained("microsoft/Phi-3-mini-128k-instruct") 
-
-    messages = [ 
-        {"role": "system", "content": "You are a helpful AI assistant."}, 
-        {"role": "user", "content": "Can you provide ways to eat combinations of bananas and dragonfruits?"}, 
-        {"role": "assistant", "content": "Sure! Here are some ways to eat bananas and dragonfruits together: 1. Banana and dragonfruit smoothie: Blend bananas and dragonfruits together with some milk and honey. 2. Banana and dragonfruit salad: Mix sliced bananas and dragonfruits together with some lemon juice and honey."}, 
-        {"role": "user", "content": msg}, 
-    ] 
-    print("pipe")
-    pipe = pipeline( 
-        "text-generation", 
-        model=model, 
-        tokenizer=tokenizer, 
-    ) 
+    input_ids = tokenizer(msg, return_tensors="pt").to("cpu")
     print("output")
- #   generation_args = { 
- #       "max_new_tokens": 500, 
- #       "return_full_text": False, 
- #       "temperature": 0.0, 
- #       "do_sample": False, 
-#    } 
-
-    output = pipe(messages) #, **generation_args) 
+    outputs = model.generate(**input_ids, max_length=500) 
     print("complete")
-    return {"message": output[0]['generated_text']}
+    return {"message": tokenizer.decode(outputs[0])}