错误描述：

File "/opt/_internal/cpython-3.7.0/lib/python3.7/site-packages/paddle/fluid/dygraph/base.py", line 277, in __impl__
return func(*args, **kwargs)
File "/opt/_internal/cpython-3.7.0/lib/python3.7/site-packages/paddle/fluid/optimizer.py", line 838, in minimize
loss, startup_program=startup_program, params_grads=params_grads)
File "/opt/_internal/cpython-3.7.0/lib/python3.7/site-packages/paddle/fluid/optimizer.py", line 752, in apply_optimize
optimize_ops = self.apply_gradients(params_grads)
File "/opt/_internal/cpython-3.7.0/lib/python3.7/site-packages/paddle/fluid/optimizer.py", line 722, in apply_gradients
optimize_ops = self._create_optimization_pass(params_grads)
File "/opt/_internal/cpython-3.7.0/lib/python3.7/site-packages/paddle/fluid/optimizer.py", line 545, in _create_optimization_pass
self._create_global_learning_rate()
File "/opt/_internal/cpython-3.7.0/lib/python3.7/site-packages/paddle/fluid/optimizer.py", line 284, in _create_global_learning_rate
"learning rate variable is create outside optimizer,"

TypeError: learning rate variable is create outside optimizer,can not create new learning rate variable for new program

情景：分布式训练中，Adam优化器上的“learning_rate"参数用fluid.layers.exponential_decay时弹出错误。单机训练不会有错误，学习率参数设置为常数也不会有错误。

相关代码：

def create_optimizer(is_distributed):
"""
Create optimizer, and decide whether to apply distributd strategy according
to the "distributed" flag
"""
l2 = fluid.regularizer.L2Decay(regularization_coeff=0.1)
optimizer = fluid.optimizer.Adam(learning_rate=fluid.layers.exponential_decay(learning_rate=args.learning_rate,decay_steps=1,
decay_rate=0.9999,
staircase=False), regularization=l2)
if is_distributed:
optimizer = distributed_optimize(optimizer)
return optimizer