Update app.py

app.py

@@ -18,6 +18,8 @@ model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.float16
 model.load_adapter(adapter_model_id)
 
 tokenizer = AutoTokenizer.from_pretrained(model_id)
+tokenizer.pad_token = tokenizer.eos_token
+tokenizer.padding_side = "right"
 
 
 def get_base_prompt():
@@ -34,19 +36,17 @@ def format_prompt(base, user_message):
 
 def predict(input, history=[]):
     input = format_prompt(get_base_prompt(), input)
-    new_user_input_ids = tokenizer.encode(f"{tokenizer.eos_token}{input}", return_tensors="pt")
+    new_user_input_ids = tokenizer.encode(f"{input}{tokenizer.eos_token}", return_tensors="pt")
     bot_input_ids = torch.cat([torch.LongTensor(history), new_user_input_ids], dim=-1)
 
     history = model.generate(
         bot_input_ids,
-        max_length=4000,
+        max_length=2000,
         pad_token_id=tokenizer.eos_token_id
     ).tolist()
 
     response = tokenizer.decode(history[0]).split("<|endoftext|>")
-    response = [
-        (response[i], response[i + 1]) for i in range(0, len(response) - 1, 2)
-    ]
+    response = [(response[i], response[i + 1]) for i in range(0, len(response) - 1, 2)]
 
     return response, history