yolo5 [Errno 32] Broken pipe OSError: [WinError 1455] 页面文件太小，无法完成 *** 作

yolo5 [Errno 32] Broken pipe OSError: [WinError 1455] 页面文件太小，无法完成 *** 作

错误代码：

  File "D:UsersAdministratorminiconda3librunpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "E:projectweilanYOLOv5-Lite-mastertrain.py", line 12, in
    import torch.distributed as dist
  File "D:UsersAdministratorminiconda3libsite-packagestorch__init__.py", line 117, in
    raise err
OSError: [WinError 1455] 页面文件太小，无法完成 *** 作。 Error loading "D:UsersAdministratorminiconda3libsite-packagestorchlibcaffe2_detectron_ops_gpu.dll" or one of its dependencies.
Traceback (most recent call last):
  File "E:/project/weilan/YOLOv5-Lite-master/train.py", line 544, in
    train(hyp, opt, device, tb_writer)
  File "E:/project/weilan/YOLOv5-Lite-master/train.py", line 203, in train
    pad=0.5, prefix=colorstr('val: '))[0]
  File "E:projectweilanYOLOv5-Lite-masterutilsdatasets.py", line 142, in create_dataloader
    collate_fn=LoadImagesAndLabels.collate_fn4 if quad else LoadImagesAndLabels.collate_fn)
  File "E:projectweilanYOLOv5-Lite-masterutilsdatasets.py", line 155, in __init__
    self.iterator = super().__iter__()
  File "D:UsersAdministratorminiconda3libsite-packagestorchutilsdatadataloader.py", line 352, in __iter__
    return self._get_iterator()
  File "D:UsersAdministratorminiconda3libsite-packagestorchutilsdatadataloader.py", line 294, in _get_iterator
    return _MultiProcessingDataLoaderIter(self)
  File "D:UsersAdministratorminiconda3libsite-packagestorchutilsdatadataloader.py", line 801, in __init__
    w.start()
  File "D:UsersAdministratorminiconda3libmultiprocessingprocess.py", line 112, in start
    self._popen = self._Popen(self)
  File "D:UsersAdministratorminiconda3libmultiprocessingcontext.py", line 223, in _Popen
    return _default_context.get_context().Process._Popen(process_obj)
  File "D:UsersAdministratorminiconda3libmultiprocessingcontext.py", line 322, in _Popen
    return Popen(process_obj)
  File "D:UsersAdministratorminiconda3libmultiprocessingpopen_spawn_win32.py", line 89, in __init__
    reduction.dump(process_obj, to_child)
  File "D:UsersAdministratorminiconda3libmultiprocessingreduction.py", line 60, in dump
    ForkingPickler(file, protocol).dump(obj)
BrokenPipeError: [Errno 32] Broken pipe

Process finished with exit code 1
 

原因：多进程加载数据，超过cpu核数了，

解决方法，减小workers，我的改为4可以了。