AttributeError: None object has no attribute: size

1. 问题说明：
    我基于mindcv仓库的[train_with_func.py](https://gitee.com/mindspore-lab/mindcv/blob/main/train_with_func.py)修改了代码，导入了Gmlp模型。修改如图所示：
    ![代码修改](https://foruda.gitee.com/images/1691935067552649260/4ef60065_7675333.jpeg "2.jpg")
                                        图 1 代码修改
2. 在ModelArts中的环境如下：
    os:欧拉
    mindspore:1.10.1
    python：3.7
    Ascend:910
    cann：
    ![输入图片说明](https://foruda.gitee.com/images/1691939333252291098/1dec1367_7675333.png "屏幕截图")

3. 两次尝试的报错
    1. 这里是在 **jupyter lab** 中通过命令：mpirun --allow-run-as-root -n 2 python3  train_with_func.py --distribute &> log.txt

**得到以下报错**

```
----------------------------------------
Num devices: 2 
Distributed mode: True 
Num training samples: 1281168
Num classes: 1000 
Num batches: 5005 
Batch size: 128 
Auto augment: autoaug 
Model: mobilenet_v2_035_224 
Model param: 73075392 
Num epochs: 300 
Optimizer: adamw 
LR: 0.001 
LR Scheduler: cosine_decay
----------------------------------------
Start training

Traceback (most recent call last):
  File "train_with_func.py", line 506, in <module>
    train(args)
  File "train_with_func.py", line 296, in train
    log_interval=args.log_interval,
  File "train_with_func.py", line 410, in train_epoch
    loss, logits = train_step(data, label)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/common/api.py", line 594, in staging_specialize
    out = _MindsporeFunctionExecutor(func, hash_obj, input_signature, process_obj, jit_config)(*args)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/common/api.py", line 98, in wrapper
    results = fn(*arg, **kwargs)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/common/api.py", line 405, in __call__
    phase = self.compile(args_list, self.fn.__name__)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/common/api.py", line 379, in compile
    is_compile = self._graph_executor.compile(self.fn, compile_args, phase, True)
AttributeError: None object has no attribute: size

----------------------------------------------------
- The Traceback of Net Construct Code:
----------------------------------------------------
The function call stack (See file '/home/ma-user/work/gmlp_mindspore/rank_1/om/analyze_fail.dat' for more details. Get instructions about `analyze_fail.dat` at https://www.mindspore.cn/search?inputValue=analyze_fail.dat):
# 0 In file train_with_func.py:389
        (loss, logits), grads = grad_fn(data, label)
                                ^
# 1 In file /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/ops/composite/base.py:554
                        return grad_(fn_, weights)(*args, **kwargs)
                               ^
# 2 In file /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/ops/composite/base.py:498
            outputs = fn(*args)
                      ^
# 3 In file train_with_func.py:373
        loss = loss_fn(logits, label)
               ^
# 4 In file /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindcv/loss/cross_entropy_smooth.py:51
        loss_logits = F.cross_entropy(
# 5 In file /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/ops/function/nn_func.py:0

----------------------------------------------------
- C++ Call Stack: (For framework developers)
----------------------------------------------------
mindspore/ccsrc/pipeline/jit/static_analysis/prim.cc:1613 GetEvaluatedValueForBuiltinTypeAttrOrMethod

Traceback (most recent call last):
  File "train_with_func.py", line 506, in <module>
    train(args)
  File "train_with_func.py", line 296, in train
    log_interval=args.log_interval,
  File "train_with_func.py", line 410, in train_epoch
    loss, logits = train_step(data, label)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/common/api.py", line 594, in staging_specialize
    out = _MindsporeFunctionExecutor(func, hash_obj, input_signature, process_obj, jit_config)(*args)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/common/api.py", line 98, in wrapper
    results = fn(*arg, **kwargs)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/common/api.py", line 405, in __call__
    phase = self.compile(args_list, self.fn.__name__)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/common/api.py", line 379, in compile
    is_compile = self._graph_executor.compile(self.fn, compile_args, phase, True)
AttributeError: None object has no attribute: size

----------------------------------------------------
- The Traceback of Net Construct Code:
----------------------------------------------------
The function call stack (See file '/home/ma-user/work/gmlp_mindspore/rank_0/om/analyze_fail.dat' for more details. Get instructions about `analyze_fail.dat` at https://www.mindspore.cn/search?inputValue=analyze_fail.dat):
# 0 In file train_with_func.py:389
        (loss, logits), grads = grad_fn(data, label)
                                ^
# 1 In file /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/ops/composite/base.py:554
                        return grad_(fn_, weights)(*args, **kwargs)
                               ^
# 2 In file /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/ops/composite/base.py:498
            outputs = fn(*args)
                      ^
# 3 In file train_with_func.py:373
        loss = loss_fn(logits, label)
               ^
# 4 In file /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindcv/loss/cross_entropy_smooth.py:51
        loss_logits = F.cross_entropy(
# 5 In file /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/ops/function/nn_func.py:0

----------------------------------------------------
- C++ Call Stack: (For framework developers)
----------------------------------------------------
mindspore/ccsrc/pipeline/jit/static_analysis/prim.cc:1613 GetEvaluatedValueForBuiltinTypeAttrOrMethod

--------------------------------------------------------------------------
Primary job  terminated normally, but 1 process returned
a non-zero exit code. Per user-direction, the job has been aborted.
--------------------------------------------------------------------------
--------------------------------------------------------------------------
mpirun detected that one or more processes exited with non-zero status, thus causing
the job to be terminated. The first process to do so was:

Process name: [[28690,1],1]
  Exit code:    1
--------------------------------------------------------------------------
```

2. 神奇的是在 **vscode** 中以相同的命令运行代码得到了更多貌似有用的报错信息

```
----------------------------------------
Num devices: 1 
Distributed mode: False 
Num training samples: 1281167
Num classes: 1000 
Num batches: 10010 
Batch size: 128 
Auto augment: autoaug 
Model: mobilenet_v2_035_224 
Model param: 73075392 
Num epochs: 300 
Optimizer: adamw 
LR: 0.001 
LR Scheduler: cosine_decay
----------------------------------------
Start training
----------------------------------------
Num devices: 1 
Distributed mode: False 
Num training samples: 1281167
Num classes: 1000 
Num batches: 10010 
Batch size: 128 
Auto augment: autoaug 
Model: mobilenet_v2_035_224 
Model param: 73075392 
Num epochs: 300 
Optimizer: adamw 
LR: 0.001 
LR Scheduler: cosine_decay
----------------------------------------
Start training
Traceback (most recent call last):
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/multiprocessing/forkserver.py", line 281, in main
    old_handlers)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/multiprocessing/forkserver.py", line 317, in _serve_one
    code = spawn._main(child_r)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/multiprocessing/spawn.py", line 114, in _main
    prepare(preparation_data)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/multiprocessing/spawn.py", line 225, in prepare
    _fixup_main_from_path(data['init_main_from_path'])
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/multiprocessing/spawn.py", line 277, in _fixup_main_from_path
    run_name="__mp_main__")
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/runpy.py", line 263, in run_path
    pkg_name=pkg_name, script_name=fname)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/runpy.py", line 96, in _run_module_code
    mod_name, mod_spec, pkg_name, script_name)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/home/ma-user/work/gmlp_mindspore/train_with_func.py", line 24, in <module>
    from gmlp import gMLPVision
  File "/home/ma-user/work/gmlp_mindspore/gmlp.py", line 115, in <module>
    class SpatialGatingUnit(nn.Cell):
  File "/home/ma-user/work/gmlp_mindspore/gmlp.py", line 121, in SpatialGatingUnit
    act=nn.Identity(),
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/nn/layer/basic.py", line 360, in __init__
    super(Identity, self).__init__()
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/nn/cell.py", line 120, in __init__
    init_pipeline()
RuntimeError: Device 1 call rtSetDevice failed, ret[507033].

----------------------------------------------------
- Ascend Error Message:
----------------------------------------------------
E39999: Inner Error, Please contact support engineer!
E39999  tsd client wait response fail, device response code[1]. unknown device error.[FUNC:WaitRsp][FILE:process_mode_manager.cpp][LINE:243]
        TraceBack (most recent call last):
        TsdOpen failed. devId=1, tdt error=31[FUNC:startAicpuExecutor][FILE:runtime.cc][LINE:1644]
        Start aicpu executor failed, retCode=0x7020009 devId=1[FUNC:DeviceRetain][FILE:runtime.cc][LINE:2093]
        Check param failed, dev can not be NULL![FUNC:PrimaryContextRetain][FILE:runtime.cc][LINE:1946]
        Check param failed, ctx can not be NULL![FUNC:PrimaryContextRetain][FILE:runtime.cc][LINE:1973]
        Check param failed, context can not be null.[FUNC:NewDevice][FILE:api_impl.cc][LINE:1315]
        new device failed, retCode=0x7010006[FUNC:SetDevice][FILE:api_impl.cc][LINE:1336]
        rtSetDevice execute failed, reason=[device retain error][FUNC:FuncErrorReason][FILE:error_message_manage.cc][LINE:49]

----------------------------------------------------
- C++ Call Stack: (For framework developers)
----------------------------------------------------
mindspore/ccsrc/plugin/device/ascend/hal/hardware/ascend_deprecated_interface.cc:293 OpenTsd

Traceback (most recent call last):
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/multiprocessing/forkserver.py", line 281, in main
    old_handlers)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/multiprocessing/forkserver.py", line 317, in _serve_one
    code = spawn._main(child_r)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/multiprocessing/spawn.py", line 114, in _main
    prepare(preparation_data)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/multiprocessing/spawn.py", line 225, in prepare
    _fixup_main_from_path(data['init_main_from_path'])
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/multiprocessing/spawn.py", line 277, in _fixup_main_from_path
    run_name="__mp_main__")
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/runpy.py", line 263, in run_path
    pkg_name=pkg_name, script_name=fname)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/runpy.py", line 96, in _run_module_code
    mod_name, mod_spec, pkg_name, script_name)
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/home/ma-user/work/gmlp_mindspore/train_with_func.py", line 24, in <module>
    from gmlp import gMLPVision
  File "/home/ma-user/work/gmlp_mindspore/gmlp.py", line 115, in <module>
    class SpatialGatingUnit(nn.Cell):
  File "/home/ma-user/work/gmlp_mindspore/gmlp.py", line 121, in SpatialGatingUnit
    act=nn.Identity(),
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/nn/layer/basic.py", line 360, in __init__
    super(Identity, self).__init__()
  File "/home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages/mindspore/nn/cell.py", line 120, in __init__
    init_pipeline()
RuntimeError: Device 1 call rtSetDevice failed, ret[507033].