在詢問了已經(jīng)提出的關于這個問題的問題之后，我繼續(xù)提出它。我試圖將字母從 A 分類到 D。所有輸入圖像都是 64x64 和灰色。我的CNN的第一層是：model = Sequential()model.add(Conv2D(32, (3,  3), input_shape = input_shape, activation = 'relu'))而且input_shape它來自何處：# Define the number of classesnum_classes = 4labels_name={'A':0,'B':1,'C':2,'D':3}img_data_list=[]labels_list=[]for dataset in data_dir_list:    img_list=os.listdir(data_path+'/'+ dataset)    print ('Loading the images of dataset-'+'{}\n'.format(dataset))    label = labels_name[dataset]    for img in img_list:    input_img=cv2.imread(data_path + '/'+ dataset + '/'+ img )          input_img=cv2.cvtColor(input_img, cv2.COLOR_BGR2GRAY)          input_img_resize=cv2.resize(input_img,(128,128))          img_data_list.append(input_img_resize)          labels_list.append(label)img_data = np.array(img_data_list)img_data = img_data.astype('float32')img_data /= 255print (img_data.shape)labels = np.array(labels_list)print(np.unique(labels,return_counts=True))#convert class labels to on-hot encodingY = np_utils.to_categorical(labels, num_classes)#Shuffle the datasetx,y = shuffle(img_data,Y, random_state=2)# Split the datasetX_train, X_test, y_train, y_test = train_test_split(x, y, test_size=0.2, random_state=2)#Defining the modelinput_shape=img_data[0].shapeprint(input_shape)