-Displaying-Configurations-in-Tensorboard.md

Displaying Configurations in TensorBoard

This is a simple demonstration of displaying training configuration in a TensorBoard, using text summaries.

Screenshot

Demo

To run the demo, run the following commands (for example):

python main.py --train-dir logs/
tensorboard --logdir logs/

If you use tb, you can launch tensorboards more easily; e.g. tb logs/*.

.gitignore

logs/
*.py[cod]
__pycache__

main.py

#
# Copyright 2017 Jongwook Choi.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#      http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

""" Example main.py for managing configuration. """

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import argparse
import tensorflow as tf


class Config(argparse.Namespace):
    """ Configuration of models and training, etc. """

    @staticmethod
    def get_parser(parser=None):
        if parser is None:
            parser = argparse.ArgumentParser(description=__doc__)
        parser.add_argument('--train-dir', default=None)
        parser.add_argument('--max-steps', default=10, type=int)
        parser.add_argument('--batch_size', default=64, type=int)
        parser.add_argument('--dataset', default='mnist', choices=['mnist', 'cifar'])
        parser.add_argument('--image-size', default='[64, 64]', type=str)
        parser.add_argument('--learning-rate', default=1e-4, type=float)
        return parser

    def __init__(self, **kwargs):
        super(Config, self).__init__(**kwargs)

        self.image_size = eval(self.image_size)
        assert isinstance(self.image_size, (tuple, list))

        # assign an random id
        import uuid
        self.train_id = uuid.uuid1().hex[:6]

        if self.train_dir is None:
            self.train_dir = "./logs/example-%s" % self.train_id
        print("Train Dir: %s" % self.train_dir)
        tf.gfile.MakeDirs(self.train_dir)

    def as_text_matrix(self):
        return [[k, str(w)] for k, w in sorted(vars(self).items())]


class Trainer(object):
    """ Main training loop of TensorFlow application. """

    def __init__(self, config):
        self.config = config
        self.summary_writer = tf.summary.FileWriter(config.train_dir)
        self.session = tf.InteractiveSession()   # FIXME: or a regular session

    def run(self):
        """ Run main training loops. """
        self.summary_op = tf.summary.merge_all()

        # Add configuration matrix as a text-summary into TensorBoard.
        config_summary = tf.summary.text('TrainConfig', tf.convert_to_tensor(self.config.as_text_matrix()), collections=[])
        self.summary_writer.add_summary(config_summary.eval(session=self.session))
        # Note: the following should be called after all text summaries are configured
        self.summary_writer.add_graph(tf.get_default_graph())

        # FIXME: The rest of code follows.
        self.session.run(tf.global_variables_initializer())
        print("[*] Training starts!")


def main():
    parser = Config.get_parser()
    args = parser.parse_args()
    config = Config(**vars(args))

    print("============== Configuration ==============")
    print(config)

    print("============== Training Starts ==============")
    Trainer(config).run()


if __name__ == '__main__':
    main()

screenshot.png