動作環境

GeForce GTX 1070 (8GB)
ASRock Z170M Pro4S [Intel Z170chipset]
Ubuntu 16.04 LTS desktop amd64
TensorFlow v1.1.0
cuDNN v5.1 for Linux
CUDA v8.0
Python 3.5.2
IPython 6.0.0 -- An enhanced Interactive Python.
gcc (Ubuntu 5.4.0-6ubuntu1~16.04.4) 5.4.0 20160609
GNU bash, version 4.3.48(1)-release (x86_64-pc-linux-gnu)

TFRecords読込み版sine curveの学習を行う過程で関連技術を学ぶ。

v0.1: http://qiita.com/7of9/items/ab27432caedbfb301650
前回: http://qiita.com/7of9/items/f8c15262c08b9562f2d9

v0.5

以下のコードでエラーが出る。

learn_sineCurve_170708.py

import numpy as np
import tensorflow as tf
import tensorflow.contrib.slim as slim

"""
v0.5 Jul. 09, 2017
  - 
v0.4 Jul. 09, 2017
  - fix bug > stops only for one epoch
    + set [num_epochs=None] for string_input_producer()
  - change parameters for shuffle_batch()
  - implement training
v0.3 Jul. 09, 2017
  - fix warning > use tf.local_variables_initializer() instead of initialize_local_variables()
  - fix warning > use tf.global_variables_initializer() instead of initialize_all_variables()
v0.2 Jul. 08, 2017
  - fix bug > OutOfRangeError (current size 0)
    + use [tf.initialize_local_variables()]
v0.1 Jul. 08, 2017
  - only read [.tfrecords]
    + add inputs_xy()
    + add read_and_decode()
"""

# codingrule: PEP8

IN_FILE = 'input_170708.tfrecords'


def read_and_decode(filename_queue):
    reader = tf.TFRecordReader()
    _, serialized_example = reader.read(filename_queue)
    features = tf.parse_single_example(
        serialized_example,
        features={
           'x_raw': tf.FixedLenFeature([], tf.string),
           'y_raw': tf.FixedLenFeature([], tf.string),
        })
    x_raw = tf.decode_raw(features['x_raw'], tf.float32)
    y_raw = tf.decode_raw(features['y_raw'], tf.float32)

    x_org = tf.reshape(x_raw, [1])
    y_org = tf.reshape(y_raw, [1])

    return x_org, y_org


def inputs_xy():
    filename = IN_FILE
    filequeue = tf.train.string_input_producer(
        [filename], num_epochs=None)

    x_org, y_org = read_and_decode(filequeue)
    return x_org, y_org

x_orgs, y_orgs = inputs_xy()
batch_size = 1  # 4
x_batch, y_batch = tf.train.shuffle_batch([x_orgs, y_orgs],
                                          batch_size,
                                          capacity=40,
                                          min_after_dequeue=batch_size)

# init_op = [tf.initialize_all_variables(), tf.initialize_local_variables()]
init_op = [tf.global_variables_initializer(), tf.local_variables_initializer()]

input_ph = tf.placeholder("float", [None,1])
output_ph = tf.placeholder("float",[None,1])

## network
hiddens = slim.stack(input_ph, slim.fully_connected, [7,7,7], 
  activation_fn=tf.nn.sigmoid, scope="hidden")
# prediction = slim.fully_connected(hiddens, 1, activation_fn=tf.nn.sigmoid, scope="output")
prediction = slim.fully_connected(hiddens, 1, activation_fn=None, scope="output")
loss = tf.contrib.losses.mean_squared_error(prediction, output_ph)

train_op = slim.learning.create_train_op(loss, tf.train.AdamOptimizer(0.001))

with tf.Session() as sess:
    sess.run(init_op)
    coord = tf.train.Coordinator()
    threads = tf.train.start_queue_runners(sess=sess, coord=coord)

    try:
        for idx in range(2):
            inpbt, outbt = sess.run([x_batch, y_batch])
            print("line87")
            _, _t_loss = sess.run([train_op, loss], 
            	                  feed_dict={input_ph:inpbt, output_ph: outbt})
            print("line90")
            if (idx + 1) % 100 == 0:
            	print("%d,%f" % (idx+1, t_loss))
    finally:
        coord.request_stop()

問題点

init_op = [tf.global_variables_initializer(), tf.local_variables_initializer()]

の実装をネットワーク構築前に書いていた。

sess.run(init_op)を実施していないので、いいかと思っていたがtrain_op定義より下にする必要があるようだ。

train_op = slim.learning.create_train_op(loss, tf.train.AdamOptimizer(0.001))

# init_op = [tf.initialize_all_variables(), tf.initialize_local_variables()]
init_op = [tf.global_variables_initializer(), tf.local_variables_initializer()]

修正後コード v0.5

learn_sineCurve_170708.py

import numpy as np
import tensorflow as tf
import tensorflow.contrib.slim as slim

"""
v0.5 Jul. 09, 2017
  - fix bug > [Attempting to use uninitialized value hidden/hidden_1/weights]
v0.4 Jul. 09, 2017
  - fix bug > stops only for one epoch
    + set [num_epochs=None] for string_input_producer()
  - change parameters for shuffle_batch()
  - implement training
v0.3 Jul. 09, 2017
  - fix warning > use tf.local_variables_initializer() instead of initialize_local_variables()
  - fix warning > use tf.global_variables_initializer() instead of initialize_all_variables()
v0.2 Jul. 08, 2017
  - fix bug > OutOfRangeError (current size 0)
    + use [tf.initialize_local_variables()]
v0.1 Jul. 08, 2017
  - only read [.tfrecords]
    + add inputs_xy()
    + add read_and_decode()
"""

# codingrule: PEP8

IN_FILE = 'input_170708.tfrecords'


def read_and_decode(filename_queue):
    reader = tf.TFRecordReader()
    _, serialized_example = reader.read(filename_queue)
    features = tf.parse_single_example(
        serialized_example,
        features={
           'x_raw': tf.FixedLenFeature([], tf.string),
           'y_raw': tf.FixedLenFeature([], tf.string),
        })
    x_raw = tf.decode_raw(features['x_raw'], tf.float32)
    y_raw = tf.decode_raw(features['y_raw'], tf.float32)

    x_org = tf.reshape(x_raw, [1])
    y_org = tf.reshape(y_raw, [1])

    return x_org, y_org


def inputs_xy():
    filename = IN_FILE
    filequeue = tf.train.string_input_producer(
        [filename], num_epochs=None)

    x_org, y_org = read_and_decode(filequeue)
    return x_org, y_org

x_orgs, y_orgs = inputs_xy()
batch_size = 1  # 4
x_batch, y_batch = tf.train.shuffle_batch([x_orgs, y_orgs],
                                          batch_size,
                                          capacity=40,
                                          min_after_dequeue=batch_size)

input_ph = tf.placeholder("float", [None,1])
output_ph = tf.placeholder("float",[None,1])

## network
hiddens = slim.stack(input_ph, slim.fully_connected, [7,7,7], 
  activation_fn=tf.nn.sigmoid, scope="hidden")
# prediction = slim.fully_connected(hiddens, 1, activation_fn=tf.nn.sigmoid, scope="output")
prediction = slim.fully_connected(hiddens, 1, activation_fn=None, scope="output")
loss = tf.contrib.losses.mean_squared_error(prediction, output_ph)

train_op = slim.learning.create_train_op(loss, tf.train.AdamOptimizer(0.001))

# init_op = [tf.initialize_all_variables(), tf.initialize_local_variables()]
init_op = [tf.global_variables_initializer(), tf.local_variables_initializer()]

with tf.Session() as sess:
    sess.run(init_op)
    coord = tf.train.Coordinator()
    threads = tf.train.start_queue_runners(sess=sess, coord=coord)

    try:
        for idx in range(30000):
            inpbt, outbt = sess.run([x_batch, y_batch])
            _, t_loss = sess.run([train_op, loss], 
            	                  feed_dict={input_ph:inpbt, output_ph: outbt})
            if (idx + 1) % 100 == 0:
            	print("%d,%f" % (idx+1, t_loss))
    finally:
        coord.request_stop()

$ python3 learn_sineCurve_170708.py
...
2017-07-09 09:01:50.796304: I tensorflow/core/common_runtime/gpu/gpu_device.cc:908] DMA: 0 
2017-07-09 09:01:50.796324: I tensorflow/core/common_runtime/gpu/gpu_device.cc:918] 0:   Y 
2017-07-09 09:01:50.796332: I tensorflow/core/common_runtime/gpu/gpu_device.cc:977] Creating TensorFlow device (/gpu:0) -> (device: 0, name: GeForce GTX 1070, pci bus id: 0000:01:00.0)
100,0.702910
200,0.413344
300,0.707156
400,0.010691
500,0.015716
600,0.030282
700,0.711572
...

動き出したが、きちんと学習しているかは未確認。