Splitting MNIST data tensorflow - tensorflow

Question

I've been following the tensorflow tutorials. I've imported the MNIST dataset and ran the code for a 2 layer convolutional neural net. It took nearly 45 minutes to train. I want to cut down the training data by discarding some of the data. How do I do that?
Here's the code:
import tensorflow as tf
import numpy as np
from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets('MNIST_data', one_hot=True)
def weight_variable(shape):
initial = tf.truncated_normal(shape, stddev=0.1)
return tf.Variable(initial)
def bias_variable(shape):
initial = tf.constant(0.1, shape=shape)
return tf.Variable(initial)
def conv2d(x, W):
return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')
def max_pool_2x2(x):
return tf.nn.max_pool(x, ksize=[1, 2, 2, 1],strides=[1, 2, 2, 1], padding='SAME')
x = tf.placeholder(tf.float32, shape=[None, 784])
y_ = tf.placeholder(tf.float32, [None, 10])
W_conv1 = weight_variable([5, 5, 1, 32])
b_conv1 = bias_variable([32])
x_image = tf.reshape(x, [-1,28,28,1])
h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1) + b_conv1)
h_pool1 = max_pool_2x2(h_conv1)
W_conv2 = weight_variable([5, 5, 32, 64])
b_conv2 = bias_variable([64])
h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2)
h_pool2 = max_pool_2x2(h_conv2)
W_fc1 = weight_variable([7 * 7 * 64, 1024])
b_fc1 = bias_variable([1024])
h_pool2_flat = tf.reshape(h_pool2, [-1, 7*7*64])
h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)
keep_prob = tf.placeholder(tf.float32)
h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)
W_fc2 = weight_variable([1024, 10])
b_fc2 = bias_variable([10])
y_conv = tf.matmul(h_fc1_drop, W_fc2) + b_fc2
cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(y_conv, y_))
train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)
correct_prediction = tf.equal(tf.argmax(y_conv,1), tf.argmax(y_,1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
sess = tf.Session()
sess.run(tf.initialize_all_variables())
for i in range(20000):
batch = mnist.train.next_batch(50)
if i%100 == 0:
train_accuracy = accuracy.eval(session=sess,feed_dict={x:batch[0], y_: batch[1], keep_prob: 1.0})
print("step %d, training accuracy %g"%(i, train_accuracy))
train_step.run(session=sess,feed_dict={x: batch[0], y_: batch[1], keep_prob: 0.5})
print("test accuracy %g"%accuracy.eval(session=sess,feed_dict={x: np.split(mnist.test.images,5)[0], y_: np.split(mnist.test.labels,5)[0], keep_prob: 1.0}))
I cut down the size of testing data since it's a numpy array. How do I do the same for training data?

Answer 1

You are using dataset provider defined in https://github.com/tensorflow/tensorflow/blob/master/tensorflow/contrib/learn/python/learn/datasets/mnist.py
To reduce number of training sample, you can change this file (Line 237), or create a modified version and use it, instead of
from tensorflow.examples.tutorials.mnist import input_data
Which points to the link that I mentioned above.

Answer 2

Cutting your training sample will not help you in any good way - as far as you use minibatches, it doesn't affect performance directly. As a better alternative, you can reduce the number of epochs and/or increase the learning rate.
Reducing data sample is a very bad idea in this case

Answer 3

just one question - we are talking about this code `https://www.tensorflow.org/tutorials/mnist/pros/ ???
So if this needs 45 minutes, i guess you are running on cpu - you should think of using gpu. I tested the code in an Azure VM N Series with a Tesla K 80 GPU and it finishes in 4 Minutes