Keras 序列模型输入形状

Question

我想训练一个基于 numpy 数组的神经网络，该数组有 4 个条目作为 X 数据，另一个数组有一个条目作为 y 数据。

X_train = [x1, x2, x3, x4]
y_train = [y1]

我想到了一件相当简单的事情，但我无法让输入形状工作。我还发现关于输入形状如何工作的信息非常少：你必须只指定 X 数据吗？ y 数据呢？

我已经尝试设置 input_dim = 4，因为这是第一个合乎逻辑的事情，但我得到了以下错误： Error when checking input: expected dense_1_input to have shape (4,) but got array with shape (1,)

然后我尝试设置 input_dim = (4, 1)，因为我认为 y 数据导致了该问题。但我又收到一条错误消息： Error when checking input: expected dense_1_input to have 3 dimensions, but got array with shape (4, 1)

代码如下：

# importing the packages
import gym
import numpy as np
from collections import deque

from keras.models import Sequential
from keras.layers import Dense
from keras.optimizers import Adam
from keras.wrappers.scikit_learn import KerasRegressor

from joblib import Parallel

# creating the environment
env = gym.make('CartPole-v1')

#defining global variables
lr=0.0001
decay=0.001
batch_size=None

# creating a deep learning model with keras
def model():
    model = Sequential()

    model.add(Dense(64, input_dim=4, activation='relu'))
    model.add(Dense(32, activation='relu'))
    model.add(Dense(16, activation='relu'))

    model.add(Dense(1, activation='sigmoid'))

    model.compile(Adam(lr=lr, decay=decay), loss='mse')
    model.summary()
    return model

# running the game
for i_episodes in range(200):
    env.reset()
    for i in range(100):
        env.render()
        action = env.action_space.sample()
        observation, reward, done, info = env.step(action)

        # observation = ndarray float64
        # reward = float
        # done = bool
        # action = int
        # info = empty

        observation = np.asarray(observation)
        reward = np.asarray(reward)
        action = np.asarray(action)

        # print(observation.dtype, reward.dtype, action.dtype)
        # print(observation.shape, action.shape)

        estimator = KerasRegressor(build_fn=model, epochs=30, batch_size=3, verbose=1)
        estimator.fit(observation, action)

        if done:
            break
env.close()

如果有人能解释输入形状的工作原理，将不胜感激。

Answer 1

兄弟！对于第二个错误，请使用此代码。现在它对我来说运行良好。

X=[]
y=[]

# creating the environment
env = gym.make('CartPole-v1')

#defining global variables
lr=0.0001
decay=0.001
batch_size=None

# creating a deep learning model with keras
def model():
    model = Sequential()

    model.add(Dense(64, input_dim=4, activation='relu'))
    model.add(Dense(32, activation='relu'))
    model.add(Dense(16, activation='relu'))

    model.add(Dense(1, activation='sigmoid'))

    model.compile(Adam(lr=lr, decay=decay), loss='mse')
    model.summary()
    return model

# running the game
for i_episodes in range(200):
    env.reset()
    for i in range(100):
        #env.render()
        action = env.action_space.sample()
        observation, reward, done, info = env.step(action)


        observation = np.asarray(observation)
        reward = np.asarray(reward)
        action = np.asarray(action)

        X.append( observation)
        y.append(action)


        if done:
            break
env.close()


X=np.asarray(X)
y=np.asarray(y)
estimator = KerasRegressor(build_fn=model, epochs=30, batch_size=3, verbose=1)
estimator.fit(X, y)

Answer 2

试试这段代码。当您要使用神经网络解决任何回归问题时，您必须指定输入维度。因此，在输入维度中，您必须将要提供给网络的列数传递给您。

  def baseline_model():  
    model = Sequential()

    model.add(Dense(64, input_dim=4, activation='relu'))
    model.add(Dense(32, activation='relu'))
    model.add(Dense(16, activation='relu'))

    model.add(Dense(1, activation='sigmoid'))

    model.compile(Adam(lr=lr, decay=decay), loss='mse')
    model.summary()
    return model

现在您必须将其包装在 keras 回归器类中，以便 keras 知道这是您要解决的回归问题。

estimator = KerasRegressor(build_fn=baseline_model, epochs=30, batch_size=3, verbose=1)

如果您需要更多关于如何使用 keras 解决回归问题的信息，而不是查看下面的我的笔记本，您将从中获得帮助。

在调用 Fit 函数之前也要使用这一行

(observation=observation.reshape(1,4))

链接：Solve Regression problem with Keras Neural Network

Answer 3

输入形状总是期望批量大小作为第一维。

例如，在您的情况下，以下图层不需要形状为 (4,) 的数组

Dense(64, input_dim=4, activation='relu')

这个密集层的输入形状是一个形状为 (n, 4) 的张量，其中 n 是批量大小。

要将您的observation 传递给模型，您首先需要按如下方式扩展其尺寸：​​

observation = np.asarray(observation)
observation = np.expand_dims(observation, axis=0) # From shape (4,) to (1, 4)
estimator.fit(observation, action)

您的代码应如下所示。

# creating a deep learning model with keras
def build_model():
    model = Sequential()

    model.add(Dense(64, input_dim=4, activation='relu'))
    model.add(Dense(32, activation='relu'))
    model.add(Dense(16, activation='relu'))

    model.add(Dense(1, activation='sigmoid'))

    model.compile(Adam(lr=lr, decay=decay), loss='mse')
    model.summary()
    return model

model = build_model()

# running the game
for i_episodes in range(200):
    env.reset()
    for i in range(100):
        env.render()
        action = env.action_space.sample()
        observation, reward, done, info = env.step(action)

        # observation = ndarray float64
        # reward = float
        # done = bool
        # action = int
        # info = empty

        observation = np.asarray(observation)
        reward = np.asarray(reward)
        action = np.asarray(action)

        model.fit(np.expand_dims(observation, axis=0), np.expand_dims(action, axis=0))

如果你正在学习 DQN，请查看article