Update app.py

app.py

@@ -90,7 +90,7 @@ with gr.Blocks() as demo:
         )
         # batch = {k: v.cuda() for k, v in batch.items()} # Using CPU only
         generated = model.generate(batch["input_ids"], max_length=1024, temperature=0.8)
-        bot_message = tokenizer.decode(generated[0][:-1]).split("### Assistant:\n", 1)[1]
+        bot_message = tokenizer.decode(generated[0][:-2]).split("### Assistant:\n", 1)[1]
 
         chat_history.append((message, bot_message))
         time.sleep(1)