Update app.py

app.py

@@ -3,20 +3,8 @@ from huggingface_hub import InferenceClient
 import requests
 import os
 
-url = "http://47.94.86.196:8085/chat_completion"
+url = "http://59.110.170.104:8085/chat_completion"
 
-"""
-class ChatRequest(BaseModel):
-    messages: typing.List[typing.Dict]
-    max_new_tokens: int
-    max_length: int = 4096
-    temperature: int = 1.0
-    top_k: int = 0
-    top_p: float = 1.0
-    repetition_penalty: float = 1.03
-    do_sample: bool = True
-    seed: int = 42
-"""
 
 def respond(
     message,
@@ -60,13 +48,15 @@ def respond(
 """
 For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
 """
+
 demo = gr.ChatInterface(
     respond,
+    chatbot=gr.Chatbot(height=600),
     additional_inputs=[
         # gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
         gr.Checkbox(True, label="do sample"),
         gr.Number(42, precision=0, label="seed"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
+        gr.Slider(minimum=1, maximum=2048, value=1024, step=1, label="Max new tokens"),
         gr.Slider(minimum=0.01, maximum=4.0, value=0.7, step=0.01, label="Temperature"),
         gr.Slider(
             minimum=0.1,
@@ -94,5 +84,5 @@ demo = gr.ChatInterface(
 
 
 if __name__ == "__main__":
-    demo.queue(default_concurrency_limit=1, max_size=5)
+    demo.queue(default_concurrency_limit=2, max_size=10)
     demo.launch()