从表格数据训练深度学习模型时，损失总是 nan

Question

我正在尝试从包含大约数千个条目的数据集中训练一个模型，该数据集具有 51 个数字特征和一个标记列，示例：

在训练模型以预测 3 个标签（候选、误报、确认）时，损失始终为 nan，并且准确度在特定值上稳定得非常快。 代码：

import tensorflow as tf
import numpy as np
import pandas as pd
import sklearn.preprocessing
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder, OneHotEncoder, StandardScaler, RobustScaler
from sklearn.preprocessing import OrdinalEncoder
from tensorflow import optimizers
from tensorflow.python.keras.layers import Dense, Dropout, Normalization
from tensorflow.python.keras.models import Sequential, Model 


def load_dataset(data_folder_csv):
    # load the dataset as a pandas DataFrame
    data = pd.read_csv(data_folder_csv, header=0)
    # retrieve numpy array
    dataset = data.values

    # split into input (X) and output (y) variables
    X = dataset[:, :-1]
    y = dataset[:, -1]
    print(y)

    # format all fields as floats
    X = X.astype(np.float)
    # reshape the output variable to be one column (e.g. a 2D shape)
    y = y.reshape((len(y), 1))
    return X, y


# prepare input data using min/max scaler.
def prepare_inputs(X_train, X_test):
    oe = RobustScaler().fit_transform(X_train)
    X_train_enc = oe.transform(X_train)
    X_test_enc = oe.transform(X_test)
    return X_train_enc, X_test_enc


# prepare target
def prepare_targets(y_train, y_test):
    le = LabelEncoder()
    ohe = OneHotEncoder()
    le.fit(y_train)
    le.fit(y_test)
    y_train_enc = ohe.fit_transform(y_train).toarray()
    y_test_enc = ohe.fit_transform(y_test).toarray()
    return y_train_enc, y_test_enc



X, y = load_dataset("csv_ready.csv")

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=1)

print('Train', X_train.shape, y_train.shape)
print('Test', X_test.shape, y_test.shape)

X_train_enc, X_test_enc = X_train, X_test
print('Finished preparing inputs.'

# prepare output data
y_train_enc, y_test_enc = prepare_targets(y_train, y_test)

norm_layer = Normalization()
norm_layer.adapt(X)
model = Sequential()
model.add(Dense(128, input_dim=X_train.shape[1], activation="tanh", kernel_initializer='he_normal'))
model.add(Dropout(0.2))
model.add(Dense(64, input_dim=X_train.shape[1], activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(32, input_dim=X_train.shape[1], activation='relu'))
model.add(Dense(3, activation='sigmoid'))

opt = optimizers.Adam(lr=0.01, decay=1e-6)
model.compile(loss='categorical_crossentropy', optimizer=opt, metrics=['accuracy'])
model.summary()

model.fit(X_train, y_train_enc, epochs=20, batch_size=128, verbose=1, use_multiprocessing=True)

_, accuracy = model.evaluate(X_test, y_test_enc, verbose=0)
print('Accuracy: %.2f' % (accuracy * 100))

我尝试增加/减少学习率，更改优化器，减少和增加神经元和层的数量，以及使用批量大小，但似乎没有任何东西可以让模型获得好的结果。我想我在这里遗漏了一些东西，但不能指望它。 结果示例：

编辑：来自 csv 的更多行：

EDIT2：也尝试了 l2 正则化，但没有做任何事情。

Answer 1

其中一个原因： 检查您的数据集是否有 NaN 值。 NaN 值可能会在学习时给模型带来问题。

您的代码中的一些主要错误：

• 对于具有 3 个神经元的输出层，您正在使用 sigmoid 激活函数而不是 softmax
• 您在使用错误的编码器时同时拟合训练集和测试集。您应该将fit_transform 用于您的训练数据，并且仅将transform 用于测试集
• 另外，您对所有层都使用了错误的输入，只有第一层应该接受输入张量。
• 您忘记为X_train 和X_test 使用prepare_inputs 函数
• 你的模型应该适合X_train_enc而不是X_train

改用这个

import tensorflow as tf
import numpy as np
import pandas as pd
import sklearn.preprocessing
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder, OneHotEncoder, StandardScaler, MinMaxScaler
from sklearn.preprocessing import OrdinalEncoder
from tensorflow import optimizers
from tensorflow.python.keras.layers import Dense, Dropout, Normalization
from tensorflow.python.keras.models import Sequential, Model 


def load_dataset(data_folder_csv):
    # load the dataset as a pandas DataFrame
    data = pd.read_csv(data_folder_csv, header=0)
    # retrieve numpy array
    dataset = data.values

    # split into input (X) and output (y) variables
    X = dataset[:, :-1]
    y = dataset[:, -1]
    print(y)

    # format all fields as floats
    X = X.astype(np.float)
    # reshape the output variable to be one column (e.g. a 2D shape)
    y = y.reshape((len(y), 1))
    return X, y

# prepare input data using min/max scaler.
def prepare_inputs(X_train, X_test):
    oe = MinMaxScaler()
    X_train_enc = oe.fit_transform(X_train)
    X_test_enc = oe.transform(X_test)
    return X_train_enc, X_test_enc

# prepare target
def prepare_targets(y_train, y_test):
    le = LabelEncoder()
    ohe = OneHotEncoder()
    y_train = le.fit_transform(y_train)
    y_test = le.transform(y_test)
    y_train_enc = ohe.fit_transform(y_train).toarray()
    y_test_enc = ohe.transform(y_test).toarray()
    return y_train_enc, y_test_enc



X, y = load_dataset("csv_ready.csv")

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=1)

print('Train', X_train.shape, y_train.shape)
print('Test', X_test.shape, y_test.shape)

#prepare_input function missing here
X_train_enc, X_test_enc = prepare_inputs(X_train, X_test)
print('Finished preparing inputs.')

# prepare output data
y_train_enc, y_test_enc = prepare_targets(y_train, y_test)

model = Sequential()
model.add(Dense(128, input_dim=X_train.shape[1], activation="relu")) 
model.add(Dropout(0.2))
model.add(Dense(128, activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(128, activation='relu'))
model.add(Dense(3, activation='softmax'))

#opt = optimizers.Adam(lr=0.01, decay=1e-6)
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
model.summary()

model.fit(X_train_enc, y_train_enc, epochs=20, batch_size=32, verbose=1, use_multiprocessing=True)

_, accuracy = model.evaluate(X_test_enc, y_test_enc, verbose=0)
print('Accuracy: %.2f' % (accuracy * 100))

Answer 2

您想将模型定义更改为：

model = Sequential()
model.add(Dense(128, input_shape=X_train.shape[1:], activation="tanh", kernel_initializer='he_normal'))
model.add(Dropout(0.2))
model.add(Dense(64, activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(32, activation='relu'))
model.add(Dense(3, activation='softmax'))

您只需要为第一层定义输入形状，Keras 会自动为后续层确定合适的形状。在定义 input_shape 时省略了批大小，这是第一个维度，因此是 input_shape=X_train.shape[1:]。

sigmoid 激活实际上会起作用（因为输出将在 0 和 1 之间变化），但您真正想要的是 softmax 激活（确保所有输出总和为 1，这是概率决定的-- 某事发生的概率是 100%，而不是 sigmoid 最终可能给你的 120%。

另外，您没有在任何地方使用您的LabelEncoder。我想你的意思是这样的：

def prepare_targets(y_train, y_test):
    le = LabelEncoder()
    ohe = OneHotEncoder()

    # teach the label encoder our labels
    le.fit(y_train)

    # turn our strings into integers
    y_train_transformed = le.transform(y_train)
    y_test_transformed = le.transform(y_test)

    # turn our integers into one-hot-encoded arrays
    y_train_enc = ohe.fit_transform(y_train_transformed).toarray()
    y_test_enc = ohe.transform(y_test_transformed).toarray()
    return y_train_enc, y_test_enc