switch readme config to gradio

README.md

@@ -3,10 +3,10 @@ title: Tricksy
 emoji: 🐠
 colorFrom: purple
 colorTo: green
-sdk: streamlit
-sdk_version: 1.29.0
+sdk: gradio
+sdk_version: 4.8.0
 app_file: app.py
-pinned: false
+pinned: true
 license: apache-2.0
 ---
 

app.py

@@ -115,4 +115,4 @@ with gr.Blocks(css=css) as iface:
             4. Applying an advanced index to a pinned tensor in PyTorch will return an unpinned copy of the indexed data, which means it needs to be recopied to pinned memory before it can be sent to the GPU. If we can override this default PyTorch behavior to allow direct CPU-GPU copying from a specified advanced index without intermediate copies, we should get a nice speedup.
         ''')
 
-iface.queue().launch(server_port=8501, show_api=False)
+iface.queue().launch(show_api=False)