Using the Speech_Paraformer-Large model reports an error OSError: [WinError -529697949] Windows Error 0xe06d7363 when running on NPU.

[Bofuser]

Describe the bug
According to the ipex-llm reference documentation, although the Speech_Paraformer-Large model can be run on the NPU and the example can be run through, when I run a longer audio file (about 25 seconds), an error is reported: OSError: [WinError -529697949] Windows Error 0xe06d7363.At the same time, it also warns and prompts[W socket.cpp:663] [c10d] The client socket has failed to connect to [WIN-U8UU51A4PON]:54791 (system error: 10049 - 在其上下文中，该请求的地址无效。)。The following is the complete output error content：

C:\Users\Admin\anaconda3\envs\FunASR_NPU\Lib\site-packages\transformers\deepspeed.py:23: FutureWarning: transformers.deepspeed module is deprecated and will be removed in a future version. Please import deepspeed modules directly from transformers.integrations
warnings.warn(
start init process group, rank: 1 world_size: 3
[W socket.cpp:663] [c10d] The client socket has failed to connect to [WIN-U8UU51A4PON]:54791 (system error: 10049 - 在其上下文中，该请求的地址无效。).
C:\Users\Admin\anaconda3\envs\FunASR_NPU\Lib\site-packages\transformers\deepspeed.py:23: FutureWarning: transformers.deepspeed module is deprecated and will be removed in a future version. Please import deepspeed modules directly from transformers.integrations
warnings.warn(
[W socket.cpp:663] [c10d] The client socket has failed to connect to [WIN-U8UU51A4PON]:54791 (system error: 10049 - 在其上下文中，该请求的地址无效。).
start init process group, rank: 2 world_size: 3
[W socket.cpp:663] [c10d] The client socket has failed to connect to [WIN-U8UU51A4PON]:54791 (system error: 10049 - 在其上下文中，该请求的地址无效。).
2025-04-16 10:04:11,898 - INFO - rank: 0, size: 3
2025-04-16 10:04:11,898 - INFO - rank: 1, size: 3
2025-04-16 10:04:11,898 - INFO - rank: 2, size: 3
0%| | 0/1 [00:00<?, ?it/s]Process Process-1:
Traceback (most recent call last):
File "C:\Users\Admin\anaconda3\envs\FunASR_NPU\Lib\multiprocessing\process.py", line 314, in _bootstrap
self.run()
File "C:\Users\Admin\anaconda3\envs\FunASR_NPU\Lib\multiprocessing\process.py", line 108, in run
self._target(*self._args, **self._kwargs)
File "C:\Users\Admin\anaconda3\envs\FunASR_NPU\Lib\site-packages\ipex_llm\transformers\npu_models\paraformer_mp.py", line 388, in run_prefill
encoder_outs = encoder_layer(
^^^^^^^^^^^^^^
File "C:\Users\Admin\anaconda3\envs\FunASR_NPU\Lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Admin\anaconda3\envs\FunASR_NPU\Lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Admin\anaconda3\envs\FunASR_NPU\Lib\site-packages\ipex_llm\transformers\npu_models\paraformer_mp.py", line 310, in forward
outputs = run_model(
^^^^^^^^^^
File "C:\Users\Admin\anaconda3\envs\FunASR_NPU\Lib\site-packages\torch\utils_contextlib.py", line 115, in decorate_context
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Admin\anaconda3\envs\FunASR_NPU\Lib\site-packages\ipex_llm\transformers\npu_models\mp_models_base.py", line 92, in run_model
_model_cache[key] = deque([backend_cls(*input_shapes) for i in range(replica)])
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Admin\anaconda3\envs\FunASR_NPU\Lib\site-packages\ipex_llm\transformers\npu_models\mp_models_base.py", line 92, in
_model_cache[key] = deque([backend_cls(*input_shapes) for i in range(replica)])
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Admin\anaconda3\envs\FunASR_NPU\Lib\site-packages\ipex_llm\transformers\npu_models\paraformer_mp.py", line 158, in init
x, mask = self.build_encoder(
^^^^^^^^^^^^^^^^^^^
File "C:\Users\Admin\anaconda3\envs\FunASR_NPU\Lib\site-packages\ipex_llm\transformers\npu_models\paraformer_mp.py", line 205, in build_encoder
x = self.self_attn_sanm(x, mask, in_feat, n_feat, n_head, fsmn_weight, qkv_bias, out_bias)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Admin\anaconda3\envs\FunASR_NPU\Lib\site-packages\ipex_llm\transformers\npu_models\mp_models_base.py", line 324, in self_attn_sanm
q_h = self.reshape(q, [b, t, h, d_k])
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Admin\anaconda3\envs\FunASR_NPU\Lib\site-packages\intel_npu_acceleration_library\backend\factory.py", line 88, in wrapper
node = fn(self, *args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Admin\anaconda3\envs\FunASR_NPU\Lib\site-packages\intel_npu_acceleration_library\backend\factory.py", line 461, in reshape
return backend_lib.reshape(self._mm, input_node, shape_node,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
OSError: [WinError -529697949] Windows Error 0xe06d7363

Looking forward to receiving your solution reply. Thank you very much.

Environment information
Using the latest lunar lake notebook.