如何使用 Tensorflow 和 Keras 解决 ValueError 问题

Question

Here's my code so far

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
import cv2 

mnist = tf.keras.datasets.mnist

(x_train, y_train), (x_test, y_test) = mnist.load_data()   # 28x28 tal fra 0-9
x_train = tf.keras.utils.normalize(x_train, axis=1).reshape(x_train.shape[0], -1)
x_test = tf.keras.utils.normalize(x_test, axis=1).reshape(x_test.shape[0], -1)

model = tf.keras.models.Sequential()
#model.add(tf.keras.layers.Flatten())   
model.add(tf.keras.layers.Dense(128, activation=tf.nn.relu, input_shape= x_train.shape[1:]))
model.add(tf.keras.layers.Dense(128, activation=tf.nn.relu))
model.add(tf.keras.layers.Dense(10, activation=tf.nn.softmax))   

model.compile(optimizer='adam',
            loss='sparse_categorical_crossentropy',
            metrics=['accuracy'])
model.fit(x_train, y_train, epochs=3)
val_loss, val_acc = model.evaluate(x_test, y_test)
print (val_loss)
print (val_acc)


model.save('projekt_tal_laeser')
new_model = tf.keras.models.load_model('projekt_tal_laeser')

predictions = new_model.predict(x_test)
print(predictions)

print(np.argmax(predictions[0]))

Then after saving that model I have then made this code

import tensorflow as tf
import numpy as np
import cv2


def prepare(filepath):
    global prediction, model
    IMG_SIZE = 28
    img_array = cv2.imread(filepath, cv2.IMREAD_GRAYSCALE)
    new_array = cv2.resize(img_array,(IMG_SIZE, IMG_SIZE))
    return new_array.reshape(-1, IMG_SIZE, IMG_SIZE, 1)

model = tf.keras.models.load_model('projekt_tal_laeser')

prediction = model.predict([prepare('number.jpg')])

print(prediction)

And the error that I am now running into with my second code is ValueError: Input 0 of layer sequential is incompatible with the layer: expected axis -1 of input shape to have value 784 but received input with shape [None, 28, 28, 1]

The reason why I'm making the second code is to make it easier to implement it into a GUI, this is my first project and any input is much appreciated:)

Answer 1

Solution:

prediction = model.predict([prepare('number.jpg').reshape(-1,784)])