Xpu上训练PPOCR_v3报错

[tzhang2014]

请提供下述完整信息以便快速定位问题/Please provide the following information to quickly locate the problem

• 系统环境/System Environment：
  Linux localhost.localdomain 4.19.90-23.32.v2101.ky10.aarch64
• 版本号/Version：
  Paddle： paddlepaddle-xpu==2.5.0
  PaddleOCR： release/2.6
  问题相关组件/Related components：
• 运行指令/Command Code：
  XPU_VISIBLE_DEVICES=0 python tools/train.py -c configs/rec/PP-OCRv3/en_PP-OCRv3_rec.yml -o
  Global.pretrained_model=./pretrain_models/en_PP-OCRv3_rec_train/best_accuracy -o Global.use_gpu=False
  Global.use_xpu=True
• 完整报错/Complete Error Message：
  Traceback (most recent call last):
  File "/root/miniforge3/envs/python38_paddle250/lib/python3.8/site-packages/multiprocess/queues.py", line 242, in _feed
  obj = _ForkingPickler.dumps(obj)
  File "/root/miniforge3/envs/python38_paddle250/lib/python3.8/site-packages/multiprocess/reduction.py", line 54, in dumps
  cls(buf, protocol, *args, **kwds).dump(obj)
  File "/root/miniforge3/envs/python38_paddle250/lib/python3.8/site-packages/dill/_dill.py", line 498, in dump
  StockPickler.dump(self, obj)
  File "/root/miniforge3/envs/python38_paddle250/lib/python3.8/pickle.py", line 487, in dump
  self.save(obj)
  File "/root/miniforge3/envs/python38_paddle250/lib/python3.8/pickle.py", line 560, in save
  f(self, obj) # Call unbound method with explicit self
  File "/root/miniforge3/envs/python38_paddle250/lib/python3.8/pickle.py", line 886, in save_tuple
  save(element)
  File "/root/miniforge3/envs/python38_paddle250/lib/python3.8/pickle.py", line 560, in save
  f(self, obj) # Call unbound method with explicit self
  File "/root/miniforge3/envs/python38_paddle250/lib/python3.8/pickle.py", line 931, in save_list
  self._batch_appends(obj)
  File "/root/miniforge3/envs/python38_paddle250/lib/python3.8/pickle.py", line 955, in _batch_appends
  save(x)
  File "/root/miniforge3/envs/python38_paddle250/lib/python3.8/pickle.py", line 578, in save
  rv = reduce(self.proto)
  RuntimeError: (PreconditionNotMet) Tensor is not in shared memory.Now only Tensor on shared memory can be serialized.
  [Hint: mmap_writer_allocation should not be null.] (at /workspace/paddle/fluid/pybind/tensor.cc:991)
  我们提供了AceIssueSolver来帮助你解答问题，你是否想要它来解答(请填写yes/no)?/We provide AceIssueSolver to solve issues, do you want it? (Please write yes/no): yes

请尽量不要包含图片在问题中/Please try to not include the image in the issue.