如何使用 Tensorboard 在同一个图上绘制不同的汇总指标？

Question

我希望能够在同一图上绘制验证集的每批次训练损失和平均验证损失张量板。当我的验证集太大而无法放入内存时，我遇到了这个问题，因此需要批处理和使用 tf.metrics 更新操作。

此问题可能适用于您希望出现在 Tensorboard 中的同一图表上的任何 Tensorflow 指标。

我可以

• 分别绘制这两个图（请参阅here）
• 将 validation-loss-per-validation-batch 绘制在与 training-loss-per-training-batch 相同的图上（验证时这没问题集可以是一个批次，我可以重复使用下面的训练摘要操作train_summ）

在下面的示例代码中，我的问题源于我的验证摘要 tf.summary.scalar 和 name=loss 被重命名为 loss_1 并因此移动到 Tensorboard 中的单独图表。据我所知，Tensorboard 采用 "same name" 并将它们绘制在同一个图表上，无论它们位于哪个文件夹中。这令人沮丧，因为 train_summ (name=loss) 只是永远写入train 文件夹和valid_summ (name=loss) 只写入valid 文件夹-但仍重命名为loss_1。

示例代码：

# View graphs with (Linux): $ tensorboard --logdir=/tmp/my_tf_model

import tensorflow as tf
import numpy as np
import os
import tempfile

def train_data_gen():
    yield np.random.normal(size=[3]), np.array([0.5, 0.5, 0.5])

def valid_data_gen():
    yield np.random.normal(size=[3]), np.array([0.8, 0.8, 0.8])

batch_size = 25
n_training_batches = 4
n_valid_batches = 2
n_epochs = 5
summary_loc = os.path.join(tempfile.gettempdir(), 'my_tf_model')
print("Summaries written to" + summary_loc)

# Dummy data
train_data = tf.data.Dataset.from_generator(train_data_gen, (tf.float32, tf.float32)).repeat().batch(batch_size)
valid_data = tf.data.Dataset.from_generator(valid_data_gen, (tf.float32, tf.float32)).repeat().batch(batch_size)
handle = tf.placeholder(tf.string, shape=[])
iterator = tf.data.Iterator.from_string_handle(handle, 
train_data.output_types, train_data.output_shapes)
batch_x, batch_y = iterator.get_next()
train_iter = train_data.make_initializable_iterator()
valid_iter = valid_data.make_initializable_iterator()

# Some ops on the data
loss = tf.losses.mean_squared_error(batch_x, batch_y)
valid_loss, valid_loss_update = tf.metrics.mean(loss)

# Write to summaries
train_summ = tf.summary.scalar('loss', loss)
valid_summ = tf.summary.scalar('loss', valid_loss)  # <- will be renamed to "loss_1"

with tf.Session() as sess:
    sess.run(tf.global_variables_initializer())
    train_handle, valid_handle = sess.run([train_iter.string_handle(), valid_iter.string_handle()])
    sess.run([train_iter.initializer, valid_iter.initializer])

    # Summary writers
    writer_train = tf.summary.FileWriter(os.path.join(summary_loc, 'train'), sess.graph)
    writer_valid = tf.summary.FileWriter(os.path.join(summary_loc, 'valid'), sess.graph)

    global_step = 0  # implicit as no actual training
    for i in range(n_epochs):
        # "Training"
        for j in range(n_training_batches):
            global_step += 1
            summ = sess.run(train_summ, feed_dict={handle: train_handle})
            writer_train.add_summary(summary=summ, global_step=global_step)
        # "Validation"
        sess.run(tf.local_variables_initializer())
        for j in range(n_valid_batches):
             _, batch_summ = sess.run([valid_loss_update, train_summ], feed_dict={handle: valid_handle})
            # The following will plot the batch loss for the validation set on the loss plot with the training data:
            # writer_valid.add_summary(summary=batch_summ, global_step=global_step + j + 1)
        summ = sess.run(valid_summ)
        writer_valid.add_summary(summary=summ, global_step=global_step)  # <- I want this on the training loss graph

我尝试过的

• 按照this issue 和this question 的建议，分离tf.summary.FileWriter 对象（一个用于训练，一个用于验证）（认为我在该问题的评论中提到了我所追求的）
• 使用tf.summary.merge 将我所有的训练和验证/测试指标合并到总体摘要操作中；做有用的簿记，但没有在同一张图表上绘制我想要的东西
• 使用 tf.summary.scalar family 属性（loss 仍会重命名为 loss_1）
• （完整的 hack 解决方案） 在 training 数据上使用 valid_loss, valid_loss_update = tf.metrics.mean(loss)，然后在每个训练批次中运行 tf.local_variables_initializer()。这确实为您提供了相同的摘要操作，因此将事物放在同一个图表上，但肯定不是您 打算 这样做吗？它也没有推广到其他指标。

上下文

• 张量流 1.9.0
• 张量板 1.9.0
• Python 3.5.2

Answer 1

Tensorboard custom_scalar plugin 是解决此问题的方法。

这里再次使用custom_scalar 将两个损失（每个训练批次 + 所有验证批次的平均值）绘制在同一图上：

# View graphs with (Linux): $ tensorboard --logdir=/tmp/my_tf_model

import os
import tempfile
import tensorflow as tf
import numpy as np
from tensorboard import summary as summary_lib
from tensorboard.plugins.custom_scalar import layout_pb2

def train_data_gen():
    yield np.random.normal(size=[3]), np.array([0.5, 0.5, 0.5])

def valid_data_gen():
    yield np.random.normal(size=[3]), np.array([0.8, 0.8, 0.8])

batch_size = 25
n_training_batches = 4
n_valid_batches = 2
n_epochs = 5
summary_loc = os.path.join(tempfile.gettempdir(), 'my_tf_model')
print("Summaries written to " + summary_loc)

# Dummy data
train_data = tf.data.Dataset.from_generator(
    train_data_gen, (tf.float32, tf.float32)).repeat().batch(batch_size)
valid_data = tf.data.Dataset.from_generator(
    valid_data_gen, (tf.float32, tf.float32)).repeat().batch(batch_size)
handle = tf.placeholder(tf.string, shape=[])
iterator = tf.data.Iterator.from_string_handle(handle, train_data.output_types,
                                               train_data.output_shapes)
batch_x, batch_y = iterator.get_next()
train_iter = train_data.make_initializable_iterator()
valid_iter = valid_data.make_initializable_iterator()

# Some ops on the data
loss = tf.losses.mean_squared_error(batch_x, batch_y)
valid_loss, valid_loss_update = tf.metrics.mean(loss)

with tf.name_scope('loss'):
    train_summ = summary_lib.scalar('training', loss)
    valid_summ = summary_lib.scalar('valid', valid_loss)

with tf.Session() as sess:
    sess.run(tf.global_variables_initializer())
    train_handle, valid_handle = sess.run([train_iter.string_handle(), valid_iter.string_handle()])
    sess.run([train_iter.initializer, valid_iter.initializer])

    writer_train = tf.summary.FileWriter(os.path.join(summary_loc, 'train'), sess.graph)
    writer_valid = tf.summary.FileWriter(os.path.join(summary_loc, 'valid'), sess.graph)

    layout_summary = summary_lib.custom_scalar_pb(
        layout_pb2.Layout(category=[
            layout_pb2.Category(
                title='losses',
                chart=[
                    layout_pb2.Chart(
                        title='losses',
                        multiline=layout_pb2.MultilineChartContent(tag=[
                            'loss/training', 'loss/valid'
                        ]))
                ])
        ]))
    writer_train.add_summary(layout_summary)

    global_step = 0
    for i in range(n_epochs):
        for j in range(n_training_batches): # "Training"
            global_step += 1
            summ = sess.run(train_summ, feed_dict={handle: train_handle})
            writer_train.add_summary(summary=summ, global_step=global_step)

        sess.run(tf.local_variables_initializer())
        for j in range(n_valid_batches):  # "Validation"
            _, batch_summ = sess.run([valid_loss_update, train_summ], feed_dict={handle: valid_handle})
        summ = sess.run(valid_summ)
        writer_valid.add_summary(summary=summ, global_step=global_step)

Here's the resulting output 在 Tensorboard 中。