数据切片问题

在cpu的mindspore r.15跑model_zoo官方faster rcnn 的ops Squeeze 算子或切片出错

【操作步骤&问题现象】

1、Tensor的 size为（2，128，4）

2、在rpn中进行 gt_bboxes_i = self.squeeze(gt_bboxes[i:i + 1:1, ::]) 或 gt_bboxes[i,::]报错 但自己test生成tensor切片不报错 

自己写的：

【截图信息】

【日志信息】（可选，上传日志内容或者附件）

Traceback (most recent call last):
  File "E:/Mindspore1.1/mindspore/model_zoo/official/cv/faster_rcnn/src/FasterRcnn/faster_rcnn_r50.py", line 514, in <module>
    x = network(x,img_metas,box,label,vaild_num)
  File "F:\py375\lib\site-packages\mindspore\nn\cell.py", line 433, in __call__
    raise err
  File "F:\py375\lib\site-packages\mindspore\nn\cell.py", line 430, in __call__
    output = self.run_construct(cast_inputs, kwargs)
  File "F:\py375\lib\site-packages\mindspore\nn\cell.py", line 352, in run_construct
    output = self.construct(*cast_inputs, **kwargs)
  File "E:/Mindspore1.1/mindspore/model_zoo/official/cv/faster_rcnn/src/FasterRcnn/faster_rcnn_r50.py", line 219, in construct
    gt_valids)
  File "F:\py375\lib\site-packages\mindspore\nn\cell.py", line 433, in __call__
    raise err
  File "F:\py375\lib\site-packages\mindspore\nn\cell.py", line 430, in __call__
    output = self.run_construct(cast_inputs, kwargs)
  File "F:\py375\lib\site-packages\mindspore\nn\cell.py", line 352, in run_construct
    output = self.construct(*cast_inputs, **kwargs)
  File "E:\Mindspore1.1\mindspore\model_zoo\official\cv\faster_rcnn\src\FasterRcnn\rpn.py", line 282, in construct
    gt_labels_i = self.squeeze(gt_labels[i:i + 1:1, ::])
  File "F:\py375\lib\site-packages\mindspore\common\tensor.py", line 248, in __getitem__
    out = tensor_operator_registry.get('__getitem__')(self, index)
  File "F:\py375\lib\site-packages\mindspore\ops\composite\multitype_ops\_compile_utils.py", line 38, in _tensor_getitem
    return tensor_index_by_tuple(self, index)
  File "F:\py375\lib\site-packages\mindspore\ops\composite\multitype_ops\_compile_utils.py", line 371, in tensor_index_by_tuple
    return _tensor_getitem_by_tuple_slice(data, tuple_index)
  File "F:\py375\lib\site-packages\mindspore\ops\composite\multitype_ops\_compile_utils.py", line 405, in _tensor_getitem_by_tuple_slice
    return P.StridedSlice(0, 0, 0, 0, shrink_axis_mask)(data, begin_strides, end_strides, step_strides)
  File "F:\py375\lib\site-packages\mindspore\ops\primitive.py", line 247, in __call__
    return _run_op(self, self.name, args)
  File "F:\py375\lib\site-packages\mindspore\common\api.py", line 77, in wrapper
    results = fn(*arg, **kwargs)
  File "F:\py375\lib\site-packages\mindspore\ops\primitive.py", line 682, in _run_op
    output = real_run_op(obj, op_name, args)
TypeError: mindspore\ccsrc\runtime\device\cpu\kernel_select_cpu.cc:219 KernelNotSupportException] Operator[StridedSlice]  input(kNumberTypeUInt8) output(kNumberTypeUInt8) is not support. This error means the current input type is not supported, please refer to the MindSpore doc for supported types.
Trace: 

针对您的问题，我这边看了下，是cpu上StridedSlice算子暂不支持切分uint8类型的Tensor。请问您这里的gt_lables是一个uint8类型的Tensor吗？ 用Ascend910跑通了