Update app.py

app.py

@@ -39,9 +39,9 @@ class LLMManager:
                 llm_instance = Llama(
                     model_path=temp_path,
                     n_ctx=MAX_CONTEXT_TOKENS,
-                    n_batch=96,
+                    n_batch=128,
                     n_threads=2,
-                    n_threads_batch=1,
+                    n_threads_batch=2,
                     use_mlock=True,
                     mmap=True,
                     low_vram=False,