我在Keras训练我的顺序模型时遇到了一些问题 . 我是这个主题的新手，因此有很多以下教程和代码片段涉及...

我正在构建一个基本的CNN，以便在正射影像的基础上区分线性基础设施和地貌特征 . 由于图像非常庞大，我不得不创建一个遵循Keras文档的数据生成器 . 该模型的编译工作正常 . 但每次我运行model.fit_generator（）命令时都会收到错误，我的一张训练图像丢失了（不是这样） . 我设置五个纪元开始，错误已经发生在第一个 .

我感谢任何可能出错的想法 .

我正在使用Ubuntu 16.04.5 LTS，带有iypthon笔记本，theano后端 .

train_path = '/path/to/train/folder'
tree_top = os.listdir(train_path)
training_filenames = os.listdir('%s/%s/' %(train_path, tree_top[0])) + os.listdir('%s/%s/' %(train_path, tree_top[1]))

valid_path = '/path/to/valid/folder'
tree_top = os.listdir(valid_path)
valid_filenames = os.listdir('%s/%s/' %(valid_path, tree_top[0])) + os.listdir('%s/%s/' %(valid_path, tree_top[1]))

数据生成器

from skimage.io import imread
from skimage.transform import resize
import numpy as np

class MY_Generator(Sequence):    # inherits from Sequence class 

    def __init__(self, image_filenames, labels, batch_size):
        self.image_filenames, self.labels = image_filenames, labels
        self.batch_size = batch_size

    def __len__(self):    # computes number of batches by dividing sample size by the batch_size
        return np.ceil(len(self.image_filenames) / float(self.batch_size))
        num_training_samples = len(self.image_filenames)
        return num_training_samples

    def __getitem__(self, idx):
        batch_x = self.image_filenames[idx * self.batch_size:(idx + 1) * self.batch_size]
        batch_y = self.labels[idx * self.batch_size:(idx + 1) * self.batch_size]

        return np.array([
            resize(imread(file_name), (200, 200))
               for file_name in batch_x]), np.array(batch_y)


my_training_batch_generator = MY_Generator(training_filenames, tree_top, batch_size)
my_validation_batch_generator = MY_Generator(valid_filenames, tree_top, batch_size)

新闻报道

model = Sequential([
        Conv2D(3, (3, 3), activation='relu', input_shape=(300,400,3)),
        Flatten(),
        Dense(2, activation='softmax'),
    ])
model.compile(Adam(lr=.0001), loss='categorical_crossentropy', metrics=['accuracy'])

model.fit_generator(generator=my_training_batch_generator,
                                          steps_per_epoch=(len(my_training_batch_generator.image_filenames) // batch_size),
                                          epochs=5,
                                          verbose=1,
                                          validation_data=my_validation_batch_generator,
                                          validation_steps=(len(my_validation_batch_generator.image_filenames) // batch_size)
                                          )

输出如下：

Epoch 1/5 --------------------------------------------- ------------------------------ IOError Traceback（最近一次调用last）in（）4 verbose = 1,5 validation_data = my_validation_batch_generator ，----> 6 validation_steps =（len（my_validation_batch_generator.image_filenames）// batch_size）7）/usr/local/lib/python2.7/dist-packages/keras/legacy/interfaces.pyc in wrapper（* args， ** kwargs）89 warnings.warn（'更新你的'object_name'调用'90'Keras 2 API：'signature，stacklevel = 2）---> 91 return func（* args，** kwargs）92 wrapper . original_function = func 93返回fit_generator中的包装器/usr/local/lib/python2.7/dist-packages/keras/engine/training.pyc（self，generator，steps_per_epoch，epochs，verbose，callbacks，validation_data，validation_steps，class_weight，max_queue_size ，workers，use_multiprocessing，shuffle，initial_epoch）1416 use_multiprocessing = use_multiprocessing，1417 shuffle = shuffle， - > 1418 initial_epoch = initial_epoch）1419 1420 @ interfaces.legacy_generator fit_generator中的methods_support /usr/local/lib/python2.7/dist-packages/keras/engine/training_generator.pyc（model，generator，steps_per_epoch，epochs，verbose，callbacks，validation_data，validation_steps，class_weight，max_queue_size，workers，use_multiprocessing， shuffle，initial_epoch）179 batch_index = 0 180而steps_done <steps_per_epoch： - > 181 generator_output = next（output_generator）182 183如果不是hasattr（generator_output，'len'）：/ usr / local / lib / python2.7 / record- packages / keras / utils / data_utils.pyc in get（self）599除了异常为e：600 self.stop（） - > 601 six.reraise（* sys.exc_info（））602 603 / usr / local / lib / get（self）593中的python2.7 / dist-packages / keras / utils / data_utils.pyc尝试：594而self.is_running（）： - > 595 inputs = self.queue.get（block = True）.get（ ）596 self.queue.task_done（）597如果输入不是None：/usr/lib/python2.7/multiprocessing/pool.pyc in get（self，timeout）565 return self._value 566 else： - > 567 raise self._value 568 569 def _set（self，i，obj）：IOErro r：[Errno 2]没有这样的文件或目录：'DJI_0168.JPG'

python keras deep-learning theano ioerror