trying to speed up the response generation

model.py

@@ -3,12 +3,13 @@ import os
 
 model_path = "llama-3.2-1B-it-Ecommerce-ChatBot-merged-F16.gguf"
 n_threads = os.cpu_count()
+print(f"number of cpu cores: {n_threads}")
 
 llm = Llama(
     model_path=model_path,
     n_ctx=512,
-    n_batch=32,
-    n_threads=n_threads,
+    n_batch=512,
+    n_threads=64,
     n_gpu_layers=-1,
     chat_format="llama-3"
 )