手头的简单任务:运行 N 个 epoch 的训练，在每个 epoch 之后计算准确的验证准确度.时期大小可以等于完整的训练集或一些预定义的迭代次数.在验证期间，每个验证集输入都必须被评估一次.

Simple task at hand: run training for N epochs performing calculating exact validation accuracy after each epoch. Epoch size can be either equal to full training set or some predefined number of iterations. During validation every validation set input has to be evaluated exactly once.

将 one_shot_iterators、可初始化的迭代器和/或该任务的句柄混合在一起的最佳方法是什么?

What would be the best way to mix together one_shot_iterators, initializable iterator and/or handle for that task?

这是我认为它应该如何工作的脚手架:

Here is scaffolding of how i see it should work:

def build_training_dataset():
    pass

def build_validation_dataset():
    pass

def construct_train_op(dataset):
    pass

def magic(iterator):
    pass

USE_CUSTOM_EPOCH_SIZE = True
CUSTOM_EPOCH_SIZE = 60
MAX_EPOCHS = 100


training_dataset = build_training_dataset()
validation_dataset = build_validation_dataset()


# Magic goes here to build a nice one-instance dataset
dataset = magic(training_dataset, validation_dataset)

train_op = construct_train_op(dataset)

# Run N epochs in which the training dataset is traversed, followed by the
# validation dataset.
with tf.Session() as sess:
    for epoch in MAX_EPOCHS:

        # train
        if USE_CUSTOM_EPOCH_SIZE:
            for _ in range(CUSTOM_EPOCH_SIZE):
                sess.run(train_op)
        else:
            while True:
                # I guess smth like this:
                try:
                    sess.run(train_op)
                except tf.errors.OutOfRangeError:
                    break # we are done with the epoch

        # validation
        validation_predictions = []
        while True:
            try:
                np.append(validation_predictions, sess.run(train_op)) # but for validation this time
            except tf.errors.OutOfRangeError:
                print('epoch %d finished with accuracy: %f' % (epoch validation_predictions.mean()))
                break 

推荐答案

由于解决方案比我预期的要复杂得多，因此它分为 2 个和平:

Since the solution is a lot messier than I expected it comes in 2 peaces:

0) 两个示例共享的辅助代码:

0) Auxiliary code shared by both examples:

USE_CUSTOM_EPOCH_SIZE = True
CUSTOM_EPOCH_SIZE = 60
MAX_EPOCHS = 100

TRAIN_SIZE = 500
VALIDATION_SIZE = 145
BATCH_SIZE = 64


def construct_train_op(batch):
    return batch


def build_train_dataset():
    return tf.data.Dataset.range(TRAIN_SIZE) 
        .map(lambda x: x + tf.random_uniform([], -10, 10, tf.int64)) 
        .batch(BATCH_SIZE)

def build_test_dataset():
    return tf.data.Dataset.range(VALIDATION_SIZE) 
        .batch(BATCH_SIZE)

1) 对于等于训练数据集大小的 epoch:

1) For epoch equal to the train dataset size:

# datasets construction
training_dataset = build_train_dataset()
validation_dataset = build_test_dataset()

# handle constructions. Handle allows us to feed data from different dataset by providing a parameter in feed_dict
handle = tf.placeholder(tf.string, shape=[])
iterator = tf.data.Iterator.from_string_handle(handle, training_dataset.output_types, training_dataset.output_shapes)
next_element = iterator.get_next()

train_op = construct_train_op(next_element)

training_iterator = training_dataset.make_initializable_iterator()
validation_iterator = validation_dataset.make_initializable_iterator()

with tf.Session() as sess:
    training_handle = sess.run(training_iterator.string_handle())
    validation_handle = sess.run(validation_iterator.string_handle())

    for epoch in range(MAX_EPOCHS):
        #train
        sess.run(training_iterator.initializer)
        total_in_train = 0
        while True:
            try:
                train_output = sess.run(train_op, feed_dict={handle: training_handle})
                total_in_train += len(train_output)
            except tf.errors.OutOfRangeError:
                assert total_in_train == TRAIN_SIZE
                break # we are done with the epoch

        # validation
        validation_predictions = []
        sess.run(validation_iterator.initializer)
        while True:
            try:
                pred = sess.run(train_op, feed_dict={handle: validation_handle})
                validation_predictions = np.append(validation_predictions, pred)
            except tf.errors.OutOfRangeError:
                assert len(validation_predictions) == VALIDATION_SIZE
                print('Epoch %d finished with accuracy: %f' % (epoch, np.mean(validation_predictions)))
                break

2) 对于自定义 epoch 大小:

2) For custom epoch size:

# datasets construction
training_dataset = build_train_dataset().repeat() # CHANGE 1
validation_dataset = build_test_dataset()

# handle constructions. Handle allows us to feed data from different dataset by providing a parameter in feed_dict
handle = tf.placeholder(tf.string, shape=[])
iterator = tf.data.Iterator.from_string_handle(handle, training_dataset.output_types, training_dataset.output_shapes)
next_element = iterator.get_next()


train_op = construct_train_op(next_element)

training_iterator = training_dataset.make_one_shot_iterator() # CHANGE 2
validation_iterator = validation_dataset.make_initializable_iterator()

with tf.Session() as sess:
    training_handle = sess.run(training_iterator.string_handle())
    validation_handle = sess.run(validation_iterator.string_handle())

    for epoch in range(MAX_EPOCHS):
        #train
        # CHANGE 3: no initiazation, not try/catch
        for _ in range(CUSTOM_EPOCH_SIZE): 
            train_output = sess.run(train_op, feed_dict={handle: training_handle})


        # validation
        validation_predictions = []
        sess.run(validation_iterator.initializer)
        while True:
            try:
                pred = sess.run(train_op, feed_dict={handle: validation_handle})
                validation_predictions = np.append(validation_predictions, pred)
            except tf.errors.OutOfRangeError:
                assert len(validation_predictions) == VALIDATION_SIZE
                print('Epoch %d finished with accuracy: %f' % (epoch, np.mean(validation_predictions)))
                break