Update app.py

app.py

@@ -216,15 +216,26 @@ def stream_iberotales_response(
         conversation.append({"role": "user", "content": user_message})
 
         prompt = tokenizer.apply_chat_template(conversation, tokenize=False, add_generation_prompt=True)
-        response = model(
-            prompt,
-            max_tokens=max_new_tokens,
-            temperature=temperature,
-            top_p=top_p,
-            top_k=top_k,
-            repeat_penalty=repetition_penalty,
-            stream=True
-        )
+        if isinstance(model, Llama):
+            response = model.create_completion(
+                prompt=prompt,
+                max_tokens=max_new_tokens,
+                temperature=temperature,
+                top_p=top_p,
+                top_k=top_k,
+                repeat_penalty=repetition_penalty,
+                stream=True
+            )
+        else:
+            response = model(
+                prompt,
+                max_tokens=max_new_tokens,
+                temperature=temperature,
+                top_p=top_p,
+                top_k=top_k,
+                repetition_penalty=repetition_penalty,
+                stream=True
+            )
 
         full_response = ""
         messages = []
@@ -382,7 +393,7 @@ custom_css = """
 """
 
 # Crear la interfaz
-with gr.Blocks(fill_height=True, title="Iberotales", css=custom_css) as demo:
+with gr.Blocks(title="Iberotales", css=custom_css) as demo:
     # Header con información del proyecto
     with gr.Row():
         with gr.Column():