我正在使用按字母的数据集制作一个土著语言翻译器。我对机器学习的了解很少，之前只做过一个两类图像分类器。最初我的代码运行得很好。我得到了分类报告和混淆矩阵。它显示了我的所有参数和不可训练的参数，以下是我的代码
from keras.preprocessing.image import ImageDataGeneratorfrom keras.models import Sequentialfrom keras.layers import Conv2D, MaxPooling2Dfrom keras.layers import Activation, Dropout, Flatten, Densefrom keras import backend as Kfrom tensorflow.keras.optimizers import Adamfrom PIL import ImageFile, Imageprint(Image.__file__)import numpyimport matplotlib.pyplot as plt# dimensions of our images.img_width, img_height = 150, 150train_data_dir = r'C:\Users\Acer\imagerec\BAYBAYIN\TRAIN'validation_data_dir = r'C:\Users\Acer\imagerec\BAYBAYIN\VAL'nb_train_samples = 51600nb_validation_samples = 12900epochs = 1batch_size = 100if K.image_data_format() == 'channels_first': input_shape = (3, img_width, img_height)else: input_shape = (img_width, img_height, 3)from keras.applications.xception import Xceptionfrom keras.models import Modelfrom keras.layers import Densevgg = Xception(include_top=False, weights='imagenet', input_shape=(), pooling='avg')x = vgg.outputx = Dense(1, activation='softmax')(x)model = Model(vgg.input, x)model.summary()model.compile(loss='binary_crossentropy', optimizer=Adam(lr=.0001), metrics=['accuracy'])# this is the augmentation configuration we will use for trainingtrain_datagen = ImageDataGenerator( rescale=1. / 255, shear_range=0.2, zoom_range=0.2, horizontal_flip=True)# this is the augmentation configuration we will use for testing:# only rescalingtest_datagen = ImageDataGenerator(rescale=1. / 255)train_generator = train_datagen.flow_from_directory( train_data_dir, target_size=(img_width, img_height), batch_size=batch_size, class_mode='binary')validation_generator = test_datagen.flow_from_directory( validation_data_dir, target_size=(img_width, img_height), batch_size=batch_size, class_mode='binary')model.fit_generator( train_generator, steps_per_epoch=nb_train_samples // batch_size, epochs=epochs, validation_data=validation_generator, validation_steps=nb_validation_samples // batch_size)print("PRINTING OUT CLASSIFICATION REPORT AND CONFUSION MATRIX")from sklearn.metrics import classification_reportfrom sklearn.metrics import confusion_matrixtest_steps_per_epoch = numpy.math.ceil(validation_generator.samples / validation_generator.batch_size)predictions = model.predict_generator(validation_generator, steps=test_steps_per_epoch)# Get most likely classpredicted_classes = numpy.argmax(predictions, axis=1)true_classes = validation_generator.classesclass_labels = list(validation_generator.class_indices.keys())report = classification_report(true_classes, predicted_classes, target_names=class_labels)print(report)cm=confusion_matrix(true_classes,predicted_classes)print(cm)plt.imshow(cm)import matplotlib.pyplot as pltimport numpy as npplt.imshow(np.random.random((48,48)), interpolation='nearest')plt.xticks(np.arange(0,48), ['A', 'BA', 'KA', 'GA', 'HA', '1', '2', '3', '4', '5', '6', '7', '8', '9', '10', '11', '12', '13', '14', '15', '16', '17', '18', '19', '20', '21', '22', '23', '24', '25', '26', '28', '29', '30', '31', '32', '33', '34', '35', '36', '37', '38', '39', '40', '41', '42', '43', '44'])plt.yticks(np.arange(0,48),['A', 'BA', 'KA', 'GA', 'HA', '1', '2', '3', '4', '5', '6', '7', '8', '9', '10', '11', '12', '13', '14', '15', '16', '17', '18', '19', '20', '21', '22', '23', '24', '25', '26', '28', '29', '30', '31', '32', '33', '34', '35', '36', '37', '38', '39', '40', '41', '42', '43', '44'])plt.show()我只是对这些特定代码行有问题
model.compile(loss='binary_crossentropy', optimizer=Adam(lr=.0001), metrics=['accuracy']) 我不知道它的作用是什么，但我怀疑应该使用分类交叉熵，因为我正在运行一个多类图像分类器，但是当我将其更改为categorical_crossentropy时，我遇到了很多错误，即使尝试使用sparse_categorical_crossentropy也是如此
有没有人知道我是否可以继续使用这些代码，或者我应该更改它，因为我的准确率也很低
回答：
model.compile 只是用来配置模型的损失函数、优化方法、损失度量、损失权重等。
你绝对应该使用分类交叉熵；你能贴出错误吗？你的模型表现不佳的原因可能有很多。
