解决EnforceNotMet: grad_op_maker_ should not be null Operator GradOpMaker has not been registered. at(

Posted 2021-08-09 叶庭云

问题描述：使用百度飞桨训练自己搭建的模型时，在运行到：
optimizer=fluid.optimizer.Adam(learning_rate=0.001)
optimizer.minimize(avg_loss)

出现了这样一个错误。如下所示：

---------------------------------------------------------------------------EnforceNotMet                             Traceback (most recent call last)<ipython-input-114-71baa0f01aff> in <module>
      1 # 配置优化器（Optimizer Function）  随机梯度下降SGD   学习率为0.001
      2 sgd_optimezer = fluid.optimizer.SGD(learning_rate=0.1)    # learning_rate为学习率
----> 3 sgd_optimezer.minimize(avg_loss)
</opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/decorator.py:decorator-gen-145> in minimize(self, loss, startup_program, parameter_list, no_grad_set, grad_clip)
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/wrapped_decorator.py in __impl__(func, *args, **kwargs)
     23     def __impl__(func, *args, **kwargs):
     24         wrapped_func = decorator_func(func)
---> 25         return wrapped_func(*args, **kwargs)
     26 
     27     return __impl__
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/base.py in __impl__(*args, **kwargs)
     85     def __impl__(*args, **kwargs):
     86         with _switch_tracer_mode_guard_(is_train=False):
---> 87             return func(*args, **kwargs)
     88 
     89     return __impl__
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/optimizer.py in minimize(self, loss, startup_program, parameter_list, no_grad_set, grad_clip)
    592             startup_program=startup_program,
    593             parameter_list=parameter_list,
--> 594             no_grad_set=no_grad_set)
    595 
    596         if grad_clip is not None and framework.in_dygraph_mode():
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/optimizer.py in backward(self, loss, startup_program, parameter_list, no_grad_set, callbacks)
    491             with program_guard(program, startup_program):
    492                 params_grads = append_backward(loss, parameter_list,
--> 493                                                no_grad_set, callbacks)
    494                 # Note: since we can't use all_reduce_op now,
    495                 #  dgc_op should be the last op of one grad.
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/backward.py in append_backward(loss, parameter_list, no_grad_set, callbacks)
    569         grad_to_var,
    570         callbacks,
--> 571         input_grad_names_set=input_grad_names_set)
    572 
    573     # Because calc_gradient may be called multiple times,
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/backward.py in _append_backward_ops_(block, ops, target_block, no_grad_dict, grad_to_var, callbacks, input_grad_names_set)
    308         # Getting op's corresponding grad_op
    309         grad_op_desc, op_grad_to_var = core.get_grad_op_desc(
--> 310             op.desc, cpt.to_text(no_grad_dict[block.idx]), grad_sub_block_list)
    311 
    312         # If input_grad_names_set is not None, extend grad_op_descs only when
EnforceNotMet: grad_op_maker_ should not be null
Operator GradOpMaker has not been registered. at [/paddle/paddle/fluid/framework/op_info.h:69]
PaddlePaddle Call Stacks: 
0       0x7efdea8b2c28p void paddle::platform::EnforceNotMet::Init<std::string>(std::string, char const*, int) + 360
1       0x7efdea8b2f77p paddle::platform::EnforceNotMet::EnforceNotMet(std::string const&, char const*, int) + 87
2       0x7efdea8b3dacp paddle::framework::OpInfo::GradOpMaker() const + 108
3       0x7efdea8ac83ep
4       0x7efdea8e1af6p
5       0x559b94463744p _PyMethodDef_RawFastCallKeywords + 596
6       0x559b94463861p _PyCFunction_FastCallKeywords + 33
7       0x559b944cf2bdp _PyEval_EvalFrameDefault + 20173
8       0x559b94413539p _PyEval_EvalCodeWithName + 761
9       0x559b94462f57p _PyFunction_FastCallKeywords + 903
10      0x559b944cb8ccp _PyEval_EvalFrameDefault + 5340
11      0x559b94413539p _PyEval_EvalCodeWithName + 761
12      0x559b94462ef5p _PyFunction_FastCallKeywords + 805
13      0x559b944ca806p _PyEval_EvalFrameDefault + 1046
14      0x559b94413539p _PyEval_EvalCodeWithName + 761
15      0x559b94462f57p _PyFunction_FastCallKeywords + 903
16      0x559b944cb8ccp _PyEval_EvalFrameDefault + 5340
17      0x559b94413539p _PyEval_EvalCodeWithName + 761
18      0x559b94414635p _PyFunction_FastCallDict + 469
19      0x559b944cc232p _PyEval_EvalFrameDefault + 7746
20      0x559b9441381ap _PyEval_EvalCodeWithName + 1498
21      0x559b94414635p _PyFunction_FastCallDict + 469
22      0x559b944cc232p _PyEval_EvalFrameDefault + 7746
23      0x559b9441381ap _PyEval_EvalCodeWithName + 1498
24      0x559b94462f57p _PyFunction_FastCallKeywords + 903
25      0x559b944ca806p _PyEval_EvalFrameDefault + 1046
26      0x559b94413539p _PyEval_EvalCodeWithName + 761
27      0x559b94462ef5p _PyFunction_FastCallKeywords + 805
28      0x559b944caa93p _PyEval_EvalFrameDefault + 1699
29      0x559b94413539p _PyEval_EvalCodeWithName + 761
30      0x559b94414424p PyEval_EvalCodeEx + 68
31      0x559b9441444cp PyEval_EvalCode + 28
32      0x559b944d9f8dp
33      0x559b944635d9p _PyMethodDef_RawFastCallKeywords + 233
34      0x559b94463861p _PyCFunction_FastCallKeywords + 33
35      0x559b944ceb94p _PyEval_EvalFrameDefault + 18340
36      0x559b9446c592p _PyGen_Send + 674
37      0x559b944cbe69p _PyEval_EvalFrameDefault + 6777
38      0x559b9446c592p _PyGen_Send + 674
39      0x559b944cbe69p _PyEval_EvalFrameDefault + 6777
40      0x559b9446c592p _PyGen_Send + 674
41      0x559b9446357dp _PyMethodDef_RawFastCallKeywords + 141
42      0x559b9446b3cfp _PyMethodDescr_FastCallKeywords + 79
43      0x559b944cf07cp _PyEval_EvalFrameDefault + 19596
44      0x559b94462ccbp _PyFunction_FastCallKeywords + 251
45      0x559b944ca806p _PyEval_EvalFrameDefault + 1046
46      0x559b94462ccbp _PyFunction_FastCallKeywords + 251
47      0x559b944caa93p _PyEval_EvalFrameDefault + 1699
48      0x559b94413539p _PyEval_EvalCodeWithName + 761
49      0x559b94414860p _PyFunction_FastCallDict + 1024
50      0x559b94432e53p _PyObject_Call_Prepend + 99
51      0x559b94425dbep PyObject_Call + 110
52      0x559b944cc232p _PyEval_EvalFrameDefault + 7746
53      0x559b9441381ap _PyEval_EvalCodeWithName + 1498
54      0x559b94462f57p _PyFunction_FastCallKeywords + 903
55      0x559b944cb8ccp _PyEval_EvalFrameDefault + 5340
56      0x559b9446c059p
57      0x559b944635d9p _PyMethodDef_RawFastCallKeywords + 233
58      0x559b94463861p _PyCFunction_FastCallKeywords + 33
59      0x559b944ceb94p _PyEval_EvalFrameDefault + 18340
60      0x559b9441381ap _PyEval_EvalCodeWithName + 1498
61      0x559b94462f57p _PyFunction_FastCallKeywords + 903
62      0x559b944caa93p _PyEval_EvalFrameDefault + 1699
63      0x559b9446c059p
64      0x559b944635d9p _PyMethodDef_RawFastCallKeywords + 233
65      0x559b94463861p _PyCFunction_FastCallKeywords + 33
66      0x559b944ceb94p _PyEval_EvalFrameDefault + 18340
67      0x559b9441381ap _PyEval_EvalCodeWithName + 1498
68      0x559b94462f57p _PyFunction_FastCallKeywords + 903
69      0x559b944ca806p _PyEval_EvalFrameDefault + 1046
70      0x559b9446c059p
71      0x559b944635d9p _PyMethodDef_RawFastCallKeywords + 233
72      0x559b94463861p _PyCFunction_FastCallKeywords + 33
73      0x559b944ceb94p _PyEval_EvalFrameDefault + 18340
74      0x559b9441381ap _PyEval_EvalCodeWithName + 1498
75      0x559b94414635p _PyFunction_FastCallDict + 469
76      0x559b94432e53p _PyObject_Call_Prepend + 99
77      0x559b94425dbep PyObject_Call + 110
78      0x559b944cc232p _PyEval_EvalFrameDefault + 7746
79      0x559b9446c43cp _PyGen_Send + 332
80      0x559b9446357dp _PyMethodDef_RawFastCallKeywords + 141
81      0x559b9446b3cfp _PyMethodDescr_FastCallKeywords + 79
82      0x559b944cf07cp _PyEval_EvalFrameDefault + 19596
83      0x559b94462ccbp _PyFunction_FastCallKeywords + 251
84      0x559b944caa93p _PyEval_EvalFrameDefault + 1699
85      0x559b9441381ap _PyEval_EvalCodeWithName + 1498
86      0x559b94414635p _PyFunction_FastCallDict + 469
87      0x559b94522b5bp
88      0x559b9446b8fbp _PyObject_FastCallKeywords + 1179
89      0x559b944cee86p _PyEval_EvalFrameDefault + 19094
90      0x559b94462ccbp _PyFunction_FastCallKeywords + 251
91      0x559b944caa93p _PyEval_EvalFrameDefault + 1699
92      0x559b94413d09p _PyEval_EvalCodeWithName + 2761
93      0x559b94462f57p _PyFunction_FastCallKeywords + 903
94      0x559b944fe383p
95      0x559b94434584p _PyMethodDef_RawFastCallDict + 404
96      0x559b944347c1p _PyCFunction_FastCallDict + 33
97      0x559b944d0007p _PyEval_EvalFrameDefault + 23575
98      0x559b94462ccbp _PyFunction_FastCallKeywords + 251
99      0x559b944caa93p _PyEval_EvalFrameDefault + 1699

# 配置优化器（Optimizer Function）  随机梯度下降SGD   学习率为0.001
# clone操作一定要在定义优化器之前进行！！！   否则也是会报错的。 
sgd_optimezer = fluid.optimizer.SGD(learning_rate=0.1)    # learning_rate为学习率
sgd_optimezer.minimize(avg_loss)

clone操作一定要在定义优化器之前进行！！！ 否则也是会报错的。 解决方法如下：

先配置优化器，再进行 clone 操作。