Update app.py

app.py

@@ -50,7 +50,7 @@ stopping_criteria_list = StoppingCriteriaList([stopping_criteria])
 
 def generate_text(prompt, context, max_length=2100):
   prompt=prompt.replace("\n", "").replace("¿","").replace("?","")
-  input_text = f'''<bos><start_of_turn>system ¿{context}?<end_of_turn><start_of_turn>user ¿{prompt}?<end_of_turn><start_of_turn>model'''
+  input_text = f'''<bos><start_of_turn>system\n{context}?<end_of_turn><start_of_turn>user\n{prompt}<end_of_turn><start_of_turn>model\n'''
   inputs = tokenizer.encode(input_text, return_tensors="pt", add_special_tokens=False).to("cuda:0")
   max_new_tokens=max_length
   generation_config = GenerationConfig(