輸入到neural.network的圖像大小異常

Question

我需要向.network輸入一張96 * 96大小的圖片，但是我得到了這個異常：

Model: "model"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_2 (InputLayer)        [(None, 96, 96, 3)]       0         
                                                                 
 sequential_1 (Sequential)   (None, 96, 96, 3)         0         
                                                                 
 rescaling (Rescaling)       (None, 96, 96, 3)         0         
                                                                 
 mobilenetv2_0.35_96 (Functi  (None, 3, 3, 1280)       410208    
 onal)                                                           
                                                                 
 flatten (Flatten)           (None, 11520)             0         
                                                                 
 dense (Dense)               (None, 15)                172815    
                                                                 
=================================================================
Total params: 583,023
Trainable params: 172,815
Non-trainable params: 410,208
_________________________________________________________________

(96, 96, 3)

ValueError: Input 0 of layer "model" is incompatible with the layer: expected shape=(None, 96, 96, 3), found shape=(32, 96, 3)

這很奇怪，當我調用image.shape()時，我得到了 (96, 96)，但是這個異常顯示這個圖像大小是 (32, 96)。

主程序

import image_load
from pathlib import Path
from tensorflow import keras

base_path = Path()
model = keras.models.load_model(base_path.cwd().joinpath("my_model"))
model.summary()
image = image_load.load_and_preprocess_image(str(base_path.joinpath("cat4.jpg")))
print(image.shape)
predict = model.predict(image)
print(predict)

加載圖像.py

import tensorflow as tf
import matplotlib.pyplot as plt


def preprocess_image(image):
    image = tf.image.decode_jpeg(image, channels=3)
    image = tf.image.resize(image, [96, 96])
    image /= 255.0  # normalize to [0,1] range

    return image


def load_and_preprocess_image(path):
    image = tf.io.read_file(path)
    return preprocess_image(image)


def show_image(path):
    plt.imshow(path)
    plt.show()

這是這張圖片：在此處輸入圖片描述

怎樣才能讓圖片順利進入model並得到預測結果呢？

Answer 1

您需要像這樣將 numpy 數組傳遞到model.predict ：

predict = model.predict(np.array([image]))[0]
print(predict)

（注意最后的 [0] 取 output 的第一個值，因為我們只傳遞了一個值的數組）