我目前正在尝试训练一个模型，我的输入管道是根据这里的答案构建的。我希望在每个epoch之后保存模型。但在训练了几个epoch后，训练就崩溃了。我了解到这是因为它将输入作为常量张量添加到图中。有人建议使用tf.placeholder来解决这个问题。不幸的是，这对我来说并没有解决问题。输入管道如下所示：

....filenames = [P_1]dataset = tf.data.TFRecordDataset(filenames)def _parse_function(example_proto):       keys_to_features = { 'data':tf.VarLenFeature(tf.float32)},       parsed_features = tf.parse_single_example(example_proto,  keys_to_features)       return tf.sparse_tensor_to_dense(parsed_features['data'])# Parse the record into tensors.dataset = dataset.map(_parse_function)# Shuffle the datasetdataset = dataset.shuffle(buffer_size=1000)# Repeat the input indefinitly dataset = dataset.repeat()      # Generate batches     dataset = dataset.batch(Batch_size) # Create a one-shot iteratoriterator = dataset.make_one_shot_iterator()data = iterator.get_next()   ....for i in range(epochs):    for ii in range(iteration):        image = sess.run(data)        ....     saver.save(sess, 'filename')

错误信息如下：

[libprotobuf FATAL external/protobuf_archive/src/google/protobuf/message_lite.cc:68] CHECK failed: (byte_size_before_serialization) == (byte_size_after_serialization): tensorflow.GraphDef was modified concurrently during serialization.terminate called after throwing an instance of 'google::protobuf::FatalException'  what():  CHECK failed: (byte_size_before_serialization) == (byte_size_after_serialization): tensorflow.GraphDef was modified concurrently during serialization.Aborted

回答：

问题似乎出在_parse_function中。请确保解析器在创建TFrecord文件时使用相同的方式。例如，确保它们具有相同的数据类型等。