mindspore将parameter转为tensor时报错

问题描述：

【操作步骤&问题现象】

1、我想要取出vgg19模型中的几个层的特征值，做一系列操作

2、我把contructz函数中的几个层保存到类的对象，通过调用类对象的属性来访问保存的特征，但是paramter不能转化为tensor

【截图信息】

 

【日志信息】（可选，上传日志内容或者附件）

File "D:\python39\lib\site-packages\mindspore\nn\cell.py", line 407, in __call__
    out = self.compile_and_run(*inputs)
  File "D:\python39\lib\site-packages\mindspore\nn\cell.py", line 734, in compile_and_run
    self.compile(*inputs)
  File "D:\python39\lib\site-packages\mindspore\nn\cell.py", line 721, in compile
    _cell_graph_executor.compile(self, *inputs, phase=self.phase, auto_parallel_mode=self._auto_parallel_mode)
  File "D:\python39\lib\site-packages\mindspore\common\api.py", line 551, in compile
    result = self._graph_executor.compile(obj, args_list, phase, use_vm, self.queue_name)
RuntimeError: mindspore\ccsrc\pipeline\jit\static_analysis\prim.cc:1220 EvalPrim] Create python object `<class 'mindspore.common.tensor.Tensor'>` failed, only support to create 'Cell' or 'Primitive' object.
The function call stack (See file 'D:\Mindspore\Mindspore_code\modelChange\output\rank_0\om/analyze_fail.dat' for more details):
# 0 In file D:\Mindspore\Mindspore_code\modelChange\output\vgg19.py(126)
        layer_2 = Tensor(module3_0_opt)
                  ^

WARNING: Logging before InitGoogleLogging() is written to STDERR
[EXCEPTION] ANALYZER(3384,1,?):2022-5-12 20:53:52 [mindspore\ccsrc\pipeline\jit\static_analysis\prim.cc:1220] EvalPrim] Create python object `<class 'mindspore.common.tensor.Tensor'>` failed, only support to create 'Cell' or 'Primitive' object.

Process finished with exit code 1

解答：

在 GRAPH_MODE 模式下返回的是 <NoneType> ，所以会出现报错的情况，你可以将 Parameter 保存起来，不用转为 Tensor。