Update app.py

app.py

@@ -80,13 +80,14 @@ def respond(
     ins_f = generate_prompt(message,None)
     inputs  =  tokenizer(ins_f, return_tensors="pt")
     input_ids = inputs["input_ids"].cuda()
+    max_new_tokens = 512
     generation_config = GenerationConfig(
             temperature=0.1,
             top_p=0.75,
             top_k=40,
             do_sample=True,
             num_beams=1,
-            max_new_tokens = 512
+            max_new_tokens = max_new_tokens
         )
 
     # Without streaming
@@ -148,4 +149,4 @@ demo = gr.ChatInterface(
 )
 if __name__ == "__main__":
     model.eval()
-    demo.launch()
+    demo.launch(share=True)