Update app.py

app.py

@@ -27,7 +27,7 @@ has_cuda = torch.cuda.is_available()
 if has_cuda:
     model = AutoModel.from_pretrained(model_name, trust_remote_code=True).cuda()  # 3.92G
 else:
-    model = AutoModel.from_pretrained(model_name, trust_remote_code=True)  # .float()
+    model = AutoModel.from_pretrained(model_name, trust_remote_code=True).float()  # .float()
 
 model = model.eval()
 
@@ -113,7 +113,7 @@ with gr.Blocks() as demo:
                 submitBtn = gr.Button("Submit", variant="primary")
         with gr.Column(scale=1):
             emptyBtn = gr.Button("Clear History")
-            max_length = gr.Slider(0, 32768, value=8192, step=1.0, label="Maximum length", interactive=True)
+            max_length = gr.Slider(0, 32768, value=8192/2, step=1.0, label="Maximum length", interactive=True)
             top_p = gr.Slider(0, 1, value=0.8, step=0.01, label="Top P", interactive=True)
             temperature = gr.Slider(0, 1, value=0.95, step=0.01, label="Temperature", interactive=True)
 
@@ -121,7 +121,7 @@ with gr.Blocks() as demo:
     past_key_values = gr.State(None)
 
     submitBtn.click(predict, [user_input, chatbot, max_length, top_p, temperature, history, past_key_values],
-                    [chatbot, history, past_key_values], show_progress=True)
+                    [chatbot, history, past_key_values], show_progress=True, api_name="predict")
     submitBtn.click(reset_user_input, [], [user_input])
 
     emptyBtn.click(reset_state, outputs=[chatbot, history, past_key_values], show_progress=True)