Model prompt change

app.py

@@ -27,22 +27,22 @@ tiny_model = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
 tokenizer = AutoTokenizer.from_pretrained(tiny_model)
 chat_model = AutoModelForCausalLM.from_pretrained(
     tiny_model,
-    torch_dtype=torch.float32)
+    torch_dtype=torch.float32,
+    device_map = "auto")
 
 pipe = pipeline(
     "text-generation",
     model=chat_model,
     tokenizer=tokenizer,
     device_map="auto",
-    max_new_tokens=200
+    max_new_tokens=100
 )
 
 def clean_chat_output(full_text):
-    # Split at the last assistant tag (TinyLlama uses <|assistant|>)
+
     if "<|assistant|>" in full_text:
         full_text = full_text.split("<|assistant|>")[-1]
 
-    # Remove any leftover "Item:" echoes
     lines = full_text.strip().split("\n")
     if lines[0].lower().startswith("item:"):
         lines = lines[1:]