diff --git a/python/llm/example/NPU/HF-Transformers-AutoModels/LLM/llama2.py b/python/llm/example/NPU/HF-Transformers-AutoModels/LLM/llama2.py
index b9c63c8c4f3..a808a551d5b 100644
--- a/python/llm/example/NPU/HF-Transformers-AutoModels/LLM/llama2.py
+++ b/python/llm/example/NPU/HF-Transformers-AutoModels/LLM/llama2.py
@@ -68,7 +68,7 @@ def get_prompt(message: str, chat_history: list[tuple[str, str]],
         trust_remote_code=True,
         attn_implementation="eager",
         load_in_low_bit="sym_int4",
-        enable_mp=True,
+        optimize_model=True,
         max_output_len=args.max_output_len,
         max_prompt_len=args.max_prompt_len,
         intra_pp=args.intra_pp,
diff --git a/python/llm/example/NPU/HF-Transformers-AutoModels/LLM/qwen2.py b/python/llm/example/NPU/HF-Transformers-AutoModels/LLM/qwen2.py
index 2aa3c263a38..fd0f1482147 100644
--- a/python/llm/example/NPU/HF-Transformers-AutoModels/LLM/qwen2.py
+++ b/python/llm/example/NPU/HF-Transformers-AutoModels/LLM/qwen2.py
@@ -55,7 +55,7 @@
         trust_remote_code=True,
         attn_implementation="eager",
         load_in_low_bit="sym_int4",
-        enable_mp=True,
+        optimize_model=True,
         max_output_len=args.max_output_len,
         max_prompt_len=args.max_prompt_len,
         intra_pp=args.intra_pp,
diff --git a/python/llm/src/ipex_llm/transformers/npu_model.py b/python/llm/src/ipex_llm/transformers/npu_model.py
index 601c7cd1aac..49bcfde1c19 100644
--- a/python/llm/src/ipex_llm/transformers/npu_model.py
+++ b/python/llm/src/ipex_llm/transformers/npu_model.py
@@ -110,16 +110,16 @@ def from_pretrained(cls, *args, **kwargs):
         ignore_argument(kwargs, "mixed_precision")
         ignore_argument(kwargs, "cpu_embedding")
         ignore_argument(kwargs, "embedding_qtype")
-        ignore_argument(kwargs, "optimize_model")
+        ignore_argument(kwargs, "enable_mp")
         ignore_argument(kwargs, "modules_to_not_convert")
         ignore_argument(kwargs, "quantization_config")
         ignore_argument(kwargs, "speculative")
         ignore_argument(kwargs, "pipeline_parallel_stages")
-        enable_mp = kwargs.pop("enable_mp", False)
+        optimize_model = kwargs.pop("optimize_model", False)
         max_output_len = kwargs.pop("max_output_len", 1024)
         max_prompt_len = kwargs.pop("max_prompt_len", max_output_len)
-        inter_pp = kwargs.pop("inter_pp", 2)
-        intra_pp = kwargs.pop("intra_pp", 2)
+        inter_pp = kwargs.pop("inter_pp", None)
+        intra_pp = kwargs.pop("intra_pp", None)
         transpose_value_cache = kwargs.pop("transpose_value_cache", True)
 
         _args = copy.deepcopy(args)
@@ -140,7 +140,7 @@ def from_pretrained(cls, *args, **kwargs):
         logger.info(f"Converting model, it may takes up to several minutes ...")
         from intel_npu_acceleration_library.compiler import create_npu_kernels
 
-        if enable_mp:
+        if optimize_model:
             invalidInputError(
                 max_prompt_len < max_output_len,
                 (
diff --git a/python/llm/src/ipex_llm/transformers/npu_models/convert_mp.py b/python/llm/src/ipex_llm/transformers/npu_models/convert_mp.py
index 7056f1f9923..3b6f4376b9c 100644
--- a/python/llm/src/ipex_llm/transformers/npu_models/convert_mp.py
+++ b/python/llm/src/ipex_llm/transformers/npu_models/convert_mp.py
@@ -28,11 +28,16 @@ def optimize_llm(
     model: torch.nn.Module,
     max_output_len=1024,
     max_prompt_len=1024,
-    inter_pp=2,
-    intra_pp=2,
+    inter_pp=None,
+    intra_pp=None,
     transpose_value_cache=True,
 ):
     if model.config.model_type == "llama":
+        if intra_pp is None:
+            intra_pp = 2
+        if inter_pp is None:
+            inter_pp = 2
+
         from ipex_llm.transformers.npu_models.llama_mp import gen_llama_fused_model_forward
         from ipex_llm.transformers.npu_models.llama_mp import DecodeRunner, PrefillRunner
         from transformers.models.llama.modeling_llama import LlamaModel
@@ -59,6 +64,11 @@ def optimize_llm(
         convert_forward(model, LlamaForCausalLM, llama2_casullm_forward)
     elif model.config.model_type == "qwen2" and model.config.intermediate_size == 8960:
         # for qwen2-1.5B
+        if intra_pp is None:
+            intra_pp = 2
+        if inter_pp is None:
+            inter_pp = 1
+
         from ipex_llm.transformers.npu_models.qwen2_mp import gen_qwen2_fused_model_forward
         from ipex_llm.transformers.npu_models.qwen2_mp import DecodeRunner, PrefillRunner
         from transformers.models.qwen2.modeling_qwen2 import Qwen2Model