Question

我正在运行SupervisedDBNregression（）模型，并希望找到该模型的最佳输入值，以使RMSE最小。我应用了粒子群优化（PSO）算法，但是我的代码给出了参数类型错误。我需要将哪个参数从float32更改为int23 / int64？预先感谢！

我尝试将设计变量更改为整数，但没有成功。

from pyswarm import pso
from dbn.tensorflow import SupervisedDBNRegression

# Objective function to be minimized
def mse(z):
  a, b, c, d, e, f , g = z
  predictedY = DBN(z).fit(trainX,trainY)
  return mean_squared_error(testY,predictedY)

# Constraint function
def DBN(z):
  a, b, c, d, e, f , g = z
  return SupervisedDBNRegression(hidden_layers_structure=[a,b],
                                    learning_rate_rbm=c,
                                    learning_rate=d,
                                    n_epochs_rbm=e,
                                    n_iter_backprop=f,
                                    batch_size=g,
                                    activation_function='relu')

constraints = [DBN]

#Define the lower and upper bounds for z
lb = [1,1,0.0000005,0.0000005, 1, 100,1]
ub = [20,20,0.1,0.1,10,5000,200]

xopt, fopt = pso(mse, lb, ub, ieqcons=constraints)

------------------错误---------------------

TypeError                                 Traceback (most recent call     last)

<ipython-input-13-19b53f82d36c> in <module>()
     24 ub = [20,20,0.1,0.1,10,5000,200]
     25 
---> 26 xopt, fopt = pso(mse, lb, ub, ieqcons=constraints)

______________________________ x帧____________________________________

/usr/local/lib/python3.6/dist-packages/dbn/tensorflow/models.py in  weight_variable(func, shape, stddev, dtype)
     22 
     23 def weight_variable(func, shape, stddev, dtype=tf.float32):
---> 24     initial = func(shape, stddev=stddev, dtype=dtype)
     25     return tf.Variable(initial)
     26 

/usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/random_ops.py in truncated_normal(shape, mean, stddev, dtype, seed, name)
    176     seed1, seed2 = random_seed.get_seed(seed)
    177     rnd = gen_random_ops.truncated_normal(
--> 178         shape_tensor, dtype, seed=seed1, seed2=seed2)
    179     mul = rnd * stddev_tensor
    180     value = math_ops.add(mul, mean_tensor, name=name)

/usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/gen_random_ops.py in truncated_normal(shape, dtype, seed, seed2, name)
    958   _, _, _op = _op_def_lib._apply_op_helper(
    959         "TruncatedNormal", shape=shape, dtype=dtype, seed=seed, seed2=seed2,
--> 960                            name=name)
    961   _result = _op.outputs[:]
    962   _inputs_flat = _op.inputs

/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/op_def_library.py in _apply_op_helper(self, op_type_name, name, **keywords)
     608               _SatisfiesTypeConstraint(base_type,
    609                                        _Attr(op_def, input_arg.type_attr),
--> 610                                        param_name=input_name)
    611             attrs[input_arg.type_attr] = attr_value
    612             inferred_from[input_arg.type_attr] = input_name

/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/op_def_library.py in _SatisfiesTypeConstraint(dtype, attr_def, param_name)
     58           "allowed values: %s" %
     59           (param_name, dtypes.as_dtype(dtype).name,
---> 60            ", ".join(dtypes.as_dtype(x).name for x in allowed_list)))
     61 
     62 

TypeError: Value passed to parameter 'shape' has DataType float32 not in list of allowed values: int32, int64

我希望PSO给出超参数z的值介于下限（lb）和上限（ub）之间。

更改哪个参数以修复类型错误并运行PSO？

0 个答案: