parser.add_argument('--max_steps', default=10000, type=int) parser.add_argument('--snapshot', default=2000, type=int, help='the frequency of saving the latest model') args = parser.parse_args() os.environ["CUDA_VISIBLE_DEVICES"] = args.gpu_id batch_size = data_loader.batch resize_height = data_loader.IMG_HEIGHT resize_width = data_loader.IMG_WIDTH depths = data_loader.IMG_CHANNELS data_shape = [batch_size, resize_height, resize_width, depths] _, _, labels_nums = create_csv_files.get_number_of_classification( args.dataset_path + 'train/') print(labels_nums) # 定义input_images为图片数据 input_images = tf.placeholder( dtype=tf.float32, shape=[None, resize_height, resize_width, depths], name='input') # 定义input_labels为labels数据 # input_labels = tf.placeholder(dtype=tf.int32, shape=[None], name='label') input_labels = tf.placeholder(dtype=tf.int32, shape=[None, labels_nums], name='label') # 定义dropout的概率 keep_prob = tf.placeholder(tf.float32, name='keep_prob')
saver = tf.train.Saver() saver.restore(sess, models_path) images_list=glob.glob(os.path.join(image_dir,'*.jpg')) for image_path in images_list: im=read_image(image_path,resize_height,resize_width,normalization=True) im=im[np.newaxis,:] #pred = sess.run(f_cls, feed_dict={x:im, keep_prob:1.0}) pre_score,pre_label = sess.run([score,class_id], feed_dict={input_images:im}) max_score=pre_score[0,pre_label] print("{} is: pre labels:{},name:{} score: {}".format(image_path,pre_label,list(labels_filename.keys())[list(labels_filename.values()).index(pre_label)], max_score)) sess.close() if __name__ == '__main__': dataset_path = './dataset/' #num of class _, dirnames, class_nums = create_csv_files.get_number_of_classification(dataset_path + 'train/') labels_filename = create_csv_files.get_classification_label(dirnames, class_nums) image_dir='./test_image' # labels_filename='dataset/label.txt' models_path='models_inception_v1/model.ckpt-10000' batch_size = 1 # resize_height = 224 # 指定存储图片高度 resize_width = 224 # 指定存储图片宽度 depths=3 data_format=[batch_size,resize_height,resize_width,depths] predict(models_path,image_dir, labels_filename, class_nums, data_format)