Update app.py

app.py

@@ -41,6 +41,9 @@ if  torch.cuda.is_available():
     modelB = AutoModelForCausalLM.from_pretrained(modelB_id, torch_dtype=torch.float16, device_map="auto")
     tokenizerB = AutoTokenizer.from_pretrained(modelB_id)
     tokenizerB.use_default_system_prompt = False
+    tokenizerB.pad_token = tokenizerB.eos_token
+
+    
 
 def make_prompt(entry):
     return  f"### Human: Don't repeat the assesments, limit to 500 words {entry} ### Assistant:"
@@ -70,8 +73,6 @@ def generate(
     if model == "A":
         model = modelA
         tokenizer = tokenizerA
-        
-
     else:
         model = modelB
         tokenizer = tokenizerB