Update app.py

app.py

@@ -20,9 +20,9 @@ DESCRIPTION = """\
 <center>👋 For more details, please check <a href=https://www.lgresearch.ai/blog/view?seq=507>our blog</a> or <a href=https://arxiv.org/abs/2412.04862>technical report</a></center>
 """
 
-MAX_MAX_NEW_TOKENS = 4096
+MAX_NEW_TOKENS = 2048
 DEFAULT_MAX_NEW_TOKENS = 512
-MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "3840"))
+MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "16384"))
 
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 
@@ -93,7 +93,7 @@ def generate(
 BOT_AVATAR = "EXAONE_logo.png"
 
 chatbot = gr.Chatbot(
-    label="EXAONE-3.0-7.8B-Instruct",
+    label="EXAONE-3.5-7.8B-Instruct",
     avatar_images=[None, BOT_AVATAR],
     layout="bubble",
     bubble_full_width=False
@@ -111,7 +111,7 @@ chat_interface = gr.ChatInterface(
         gr.Slider(
             label="Max new tokens",
             minimum=1,
-            maximum=MAX_MAX_NEW_TOKENS,
+            maximum=MAX_NEW_TOKENS,
             step=1,
             value=DEFAULT_MAX_NEW_TOKENS,
         ),