adapt chatbot

app.py

@@ -6,7 +6,7 @@ import gradio as gr
 model_id = "Qwen/Qwen3-1.7B"
 model = AutoModelForCausalLM.from_pretrained(
     model_id,
-    torch_dtype = torch.bfloat16
+    torch_dtype = "auto"
 )
 tokenizer = AutoTokenizer.from_pretrained(model_id)
 
@@ -38,17 +38,15 @@ def respuesta(
 
     input_ids = tokenizer.apply_chat_template(
         messages,
+        tokenize=False,
         add_generation_prompt=True,
-        return_tensors='pt'
+        enable_thinking=True
     ).to(model.device)
 
-    # terminators = [
-    #     tokenizer.eos_token_id,
-    #     tokenizer.convert_tokens_to_ids("<|eot_id|>")
-    # ]
+    model_inputs = tokenizer([text], return_tensor='pt')
 
     outputs = model.generate(
-        input_ids,
+        **model_inputs,
         max_new_tokens=max_tokens,
         do_sample=True,
         temperature=temperature,