Environment

• Hardware Environment(Ascend/GPU/CPU): GPU

Uncomment only one /device <> line, hit enter to put that in a new line, and remove leading whitespaces from that line:
/device gpu

• Software Environment:
  -- MindSpore version (source or binary):1.5.0 conda
  -- Python version (e.g., Python 3.7.5):Python 3.7.5
  -- OS platform and distribution (e.g., Linux Ubuntu 16.04):ubuntu18.04
  -- GCC/Compiler version (if compiled from source): GCC 7.5.0

Related testcase

Steps to reproduce the issue

1. 启动训练命令为：
   bash run_standalone_train_gpu.sh /path/to/cifar-10-batches-bin /path/to/resnet/config/resnet50_cifar10_config.yaml
   mindspore/models/official/cv/resnet/config/
   训练可以正常进行。
2. 如果使用简易模式启动boost，将"resnet50_cifar10_config.yaml” 中
   “boost_mode”的值由 "O0”改为"O1”或"O2”,单GPU启动训练，程序报错：
   “RuntimeError: mindspore/ccsrc/runtime/device/gpu/kernel_info_setter.cc:83 SupportedTypeList] Unsupported op [Centralization] on GPU

In file /home/dasblue/anaconda3/envs/msgpu/lib/python3.7/site-packages/mindspore/boost/less_batch_normalization.py(81)
x = self.multiplier * x
^”
3.然而使用高阶模式使用boost执行LessBN，训练是OK的。

Describe the current behavior

Describe the expected behavior

Related log / screenshot

Special notes for this issue