Adjust queue size

app.py

@@ -377,8 +377,8 @@ with gr.Blocks(title="ChatGLM2-6B-int4", theme=gr.themes.Soft(text_size="sm")) a
 # demo.queue().launch(share=True, inbrowser=True, debug=True)
 
 # concurrency_count > 1 requires more memory, max_size: queue size
-# T4 medium: 30GB, model size: ~6G concurrency_count = 3
+# T4 medium: 30GB, model size: ~4G concurrency_count = 6
 # leave one for api access
-# reduce to 3 if OOM
+# reduce to 5 if OOM occurs to often
 
-demo.queue(concurrency_count=4, max_size=20).launch(debug=True)
+demo.queue(concurrency_count=6, max_size=30).launch(debug=True)