Update app.py

app.py

@@ -70,12 +70,13 @@ stopping_criteria_list = StoppingCriteriaList([stopping_criteria])
 
 def generate_text(modelin, prompt, context, max_length=2100):
   print('Modelo es: '+modelin)
-  sft_model = modelin
-  ft_model = PeftModel.from_pretrained(base_model, sft_model)
-  model = ft_model.merge_and_unload()
+  #sft_model = modelin
+  #ft_model = PeftModel.from_pretrained(base_model, sft_model)
+  #model = ft_model.merge_and_unload()
 
   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 ¿{context}?<end_of_turn><start_of_turn>user ¿{prompt}?<end_of_turn><start_of_turn>model'''
+  input_text = str(context)+str(prompt)
   inputs = tokenizer.encode(input_text, return_tensors="pt", add_special_tokens=False).to("cuda:0")
   max_new_tokens=max_length
   generation_config = GenerationConfig(