ValueError:检查输入时出错：预期密集_1_输入具有形状（8），但获得具有形状（1，）的数组

Question

我试图训练一个神经网络来弹起一个球，但我在预测球的运动时遇到了问题，得到了错误ValueError: Error when checking input: expected dense_1_input to have shape (8,) but got array with shape (1,)
我的代码：

from keras.models import Sequential
from keras.layers import Dense
import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
import csv
import pygame

running=True

def main():
    # load training data
    data_path = 'x.dat'
    with open(data_path, 'r') as f:
        reader = csv.reader(f, delimiter=',')
        headers = next(reader)
        x_train = np.array(list(reader)).astype(float)
    data_path = 'y.dat'
    with open(data_path, 'r') as f:
        reader = csv.reader(f, delimiter=',')
        headers = next(reader)
        y_train = np.array(list(reader)).astype(float)
    
    # debug print statement
    print(x_train)
    
    # define the keras model
    model = Sequential()
    model.add(Dense(8, input_shape=(8,), activation='relu'))
    model.add(Dense(10, activation='relu'))
    model.add(Dense(4, activation='relu'))

    # compile the keras model
    model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])
    
    # fit the keras model on the dataset
    model.fit(x_train, y_train, epochs=1, batch_size=1000)

    # pygame initialization to visualize the ball
    global running, screen

    pygame.init()
    screen = pygame.display.set_mode((200, 200))
    pygame.display.set_caption("BallPhysics")
    screen.fill((255,255,255))
    pygame.display.update()

    # ball info
    x = 25
    y = 10
    xVel = -2
    yVel = 0
    gravity = 0.1
    elasticity = 0.9999
    radius = 10
    friction = 0.999

    while running:
        screen.fill((255,255,255))
        ev = pygame.event.get()

        # draw ball
        pygame.draw.circle(screen, (255, 0, 0), (x,y), radius)
        pygame.display.update()

        # make input array
        inp = [x/200,y/200,((xVel/10)+1)/2,((yVel/10)+1)/2,gravity,elasticity,radius/50.0,friction]
        print(inp)
        out = model.predict(inp)

        # set ball position and velocity to NN output
        x = out[0][0]
        y = out[0][1]
        xVel = out[0][2]
        yVel = out[0][3]
        
        # event handling
        for event in ev:
            if event.type == pygame.QUIT:
                running = False
                pygame.display.quit()
                pygame.quit()
        
            pygame.display.flip()

main()

调试打印语句将打印出来

[[0.025568 0.131659 0.755605 ... 0.414219 0.094692 0.678865]
...
[0.08742  0.08742  0.5      ... 0.250432 0.699359 0.179118]]

这有点让人困惑，因为我注意到打印出来的数组没有逗号，而我制作的数组有逗号。这可能与此有关，但我不知道是什么原因。 感谢您的帮助

堆栈跟踪：

Traceback (most recent call last):
  File "/Users/grimtin10/Documents/Python Projects/BallPhysics/BallPhysics.py", line 78, in <module>
    main()
  File "/Users/grimtin10/Documents/Python Projects/BallPhysics/BallPhysics.py", line 63, in main
    out = model.predict(inp)
  File "/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/keras/engine/training.py", line 1149, in predict
    x, _, _ = self._standardize_user_data(x)
  File "/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/keras/engine/training.py", line 751, in _standardize_user_data
    exception_prefix='input')
  File "/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/keras/engine/training_utils.py", line 138, in standardize_input_data
    str(data_shape))
ValueError: Error when checking input: expected dense_1_input to have shape (8,) but got array with shape (1,)

Answer 1

好吧，我想起来了。 事实证明，NN预期的是批处理输入，导致单个数组无法工作。我对tf和keras不太熟悉，所以这就是问题所在。 使用np.reshape(inp,(1,8))修复了它