Update app.py

app.py

@@ -77,7 +77,7 @@ def respond(
         if llama == None:
             llama = Llama("models/madlad400-3b-mt-q8_0.gguf",flash_attn=False,
                         n_gpu_layers=0,
-                        n_batch=16,
+                        n_batch=32,
                         n_ctx=512,
                         n_threads=2,
                         n_threads_batch=2,verbose=False)