模型导入In [6]:model = tf.keras.models.load_model('znxlsj003.h5')4 设计预测predict_fun()函数In [9]:def predict_fun(img): img = tf.expand_dims(img,0) return model.predict(img)5 用predict_fun()函数预测指定的图片In [10]:y1=predict_fun(test_image[1])y2=predict_fun(test_image[222])y3=predict_fun(test_image[333])1/1 [==============================] - 0s 230ms/step1/1 [==============================] - 0s 20ms/step1/1 [==============================] - 0s 28ms/stepIn [11]:print(class_label[np.argmax(y1)])print(class_label[np.argmax(y2)])print(class_label[np.argmax(y3)])shiptruckfrogIn [ ]: