Update app.py

app.py

@@ -52,7 +52,7 @@ def load_models():
     
     app.state.deepseek_model = AutoModelForCausalLM.from_pretrained(
         model_name,
-        max_tokens=10,  # Limit to very few tokens
+        # max_tokens=10,  # Limit to very few tokens
         temperature=0.1,    # Low temperature for more deterministic output
         do_sample=False,    # Disable sampling for consistent results
         torch_dtype="auto",