标签: hyperopt

我正在使用 Python 的 hyperopt 库来执行 ML 超参数的优化。特别是我试图使用这个函数来找到 lightgbm 最优超参数来最小化：

def lgb_objective_map(params):
"""
objective function for lightgbm using MAP as success metric.
"""

# hyperopt casts as float
params['num_boost_round'] = int(params['num_boost_round'])
params['num_leaves'] = int(params['num_leaves'])
params['min_data_in_leaf'] = int(params['min_data_in_leaf'])

# need to be passed as parameter
params['verbose'] = -1
params['seed'] = 1

# Cross validation
cv_result = lgb.cv(
params,
lgtrain,
nfold=3,
metrics='binary_logloss',
num_boost_round=params['num_boost_round'],
early_stopping_rounds=20,
stratified=False,
)

# Update the number of trees based on the early stopping results
early_stop_dict[lgb_objective_map.i] = len(cv_result['binary_logloss-mean'])
params['num_boost_round'] = …

我正在使用 hyperopt 来调整我的 ML 模型，但在使用 qloguniform 作为搜索空间时遇到了麻烦。我给出了来自官方维基的例子并改变了搜索空间。

import pickle
import time
#utf8
import pandas as pd
import numpy as np
from hyperopt import fmin, tpe, hp, STATUS_OK, Trials

def objective(x):
    return {
        'loss': x ** 2,
        'status': STATUS_OK,
        # -- store other results like this
        'eval_time': time.time(),
        'other_stuff': {'type': None, 'value': [0, 1, 2]},
        # -- attachments are handled differently
        'attachments':
            {'time_module': pickle.dumps(time.time)}
        }
trials = Trials()
best = fmin(objective,
    space=hp.qloguniform('x', np.log(0.001), np.log(0.1), np.log(0.001)),
    algo=tpe.suggest,
    max_evals=100,
    trials=trials)
pd.DataFrame(trials.trials) …

我使用hyperopt来搜索SVM分类器的最佳参数,但Hyperopt说最好的'内核'是'0'.{'kernel':'0'}显然不合适.

有谁知道这是由我的错误还是一袋hyperopt造成的？

代码如下.

from hyperopt import fmin, tpe, hp, rand
import numpy as np
from sklearn.metrics import accuracy_score
from sklearn import svm
from sklearn.cross_validation import StratifiedKFold

parameter_space_svc = {
   'C':hp.loguniform("C", np.log(1), np.log(100)),
   'kernel':hp.choice('kernel',['rbf','poly']),
   'gamma': hp.loguniform("gamma", np.log(0.001), np.log(0.1)),    
}

from sklearn import datasets
iris = datasets.load_digits()

train_data = iris.data
train_target = iris.target

count = 0

def function(args):
  print(args)
  score_avg = 0
  skf = StratifiedKFold(train_target, n_folds=3, shuffle=True, random_state=1)
  for train_idx, test_idx in skf:
    train_X = iris.data[train_idx]
    train_y = iris.target[train_idx]
    test_X = iris.data[test_idx] …

我使用hyperopt优化了我的keras模型。现在，我们如何将最佳优化的keras模型及其权重保存到磁盘。

我的代码：

from hyperopt import fmin, tpe, hp, STATUS_OK, Trials
from sklearn.metrics import roc_auc_score
import sys

X = []
y = []
X_val = []
y_val = []

space = {'choice': hp.choice('num_layers',
                    [ {'layers':'two', },
                    {'layers':'three',
                    'units3': hp.uniform('units3', 64,1024), 
                    'dropout3': hp.uniform('dropout3', .25,.75)}
                    ]),

            'units1': hp.choice('units1', [64,1024]),
            'units2': hp.choice('units2', [64,1024]),

            'dropout1': hp.uniform('dropout1', .25,.75),
            'dropout2': hp.uniform('dropout2',  .25,.75),

            'batch_size' : hp.uniform('batch_size', 20,100),

            'nb_epochs' :  100,
            'optimizer': hp.choice('optimizer',['adadelta','adam','rmsprop']),
            'activation': 'relu'
        }

def f_nn(params):   
    from keras.models import Sequential
    from keras.layers.core import Dense, Dropout, Activation
    from …

我尝试将贝叶斯优化应用于 MNIST 手写数字数据集的简单 CNN，但几乎没有迹象表明它有效。我已经尝试进行 k 折验证以消除噪声，但似乎优化仍然没有在收敛到最佳参数方面取得任何进展。一般来说，贝叶斯优化可能失败的一些主要原因是什么？在我的特殊情况下？

其余的只是上下文和代码片段。

型号定义：

def define_model(learning_rate, momentum):
    model = Sequential()
    model.add(Conv2D(32, (3,3), activation = 'relu', kernel_initializer = 'he_uniform', input_shape=(28,28,1)))
    model.add(MaxPooling2D((2,2)))
    model.add(Flatten())
    model.add(Dense(100, activation='relu', kernel_initializer='he_uniform'))
    model.add(Dense(10, activation='softmax'))
    opt = SGD(lr=learning_rate, momentum=momentum)
    model.compile(optimizer=opt, loss='categorical_crossentropy', metrics=['accuracy'])
    return model

使用超参数运行一次训练：batch_size = 32，学习率 = 1e-2，Momentum = 0.9，10 个 epoch。（蓝色 = 训练，黄色 = 验证）。

space = …