with tf.name_scope('b_fc2'):
b_fc2 = bias_variable([2])
tf.summary.histogram('fc_layer2', b_fc2)
with tf.name_scope('y_conv'):
y_conv = tf.matmul(h_fc1_drop, W_fc2) + b_fc2
tf.summary.histogram('fc_layer2', y_conv)
return y_conv, keep_prob
# Import data
num_classes = 2
swallowsound = read_data_sets(dir,
gzip_compress=True,
train_imgaes='train-images-idx3-ubyte.gz',
train_labels='train-labels-idx1-ubyte.gz',
test_imgaes='t10k-images-idx3-ubyte.gz',
test_labels='t10k-labels-idx1-ubyte.gz',
one_hot=True,
validation_size=50,
num_classes=num_classes,
MSB=True)
# Create the model
with tf.name_scope('inputs'):
x = tf.placeholder(tf.float32, [None, 250], name='x_input')
# Define loss and optimizer
y_ = tf.placeholder(tf.float32, [None, 2], name='y_input')
# Build the graph for the deep net
y_conv, keep_prob = deepnn(x)
# cross_entropy = tf.nn.softmax_cross_entropy_with_logits(labels=y_,logits=y_conv)
def main(_):
# Import data
num_classes = 2
swallowsound = read_data_sets(
FLAGS.data_dir,
gzip_compress=False,
train_imgaes='train-images-idx3-ubyte',
train_labels='train-labels-idx1-ubyte',
test_imgaes='t10k-images-idx3-ubyte',
test_labels='t10k-labels-idx1-ubyte',
one_hot=True,
validation_size=2000, #验证集大小
num_classes=num_classes,
MSB=True)
# Create the model
x = tf.placeholder(tf.float32, [None, 50])
# Define loss and optimizer
y_ = tf.placeholder(tf.float32, [None, 2])
# Build the graph for the deep net
y_conv, keep_prob = deepnn(x)
with tf.name_scope('loss'):
cross_entropy = tf.nn.softmax_cross_entropy_with_logits(labels=y_,
logits=y_conv)
cross_entropy = tf.reduce_mean(cross_entropy)
with tf.name_scope('adam_optimizer'):
train_step = tf.train.GradientDescentOptimizer(0.01).minimize(
cross_entropy) #accuracy 0.97
#train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy) #accuracy 0.94
with tf.name_scope('accuracy'):
correct_prediction = tf.equal(tf.argmax(y_conv, 1), tf.argmax(y_, 1))
correct_prediction = tf.cast(correct_prediction, tf.float32)
accuracy = tf.reduce_mean(correct_prediction)
graph_location = tempfile.mkdtemp()
print('Saving graph to: %s' % graph_location)
train_writer = tf.summary.FileWriter(graph_location)
train_writer.add_graph(tf.get_default_graph())
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
for i in range(10000):
batch = swallowsound.train.next_batch(
1000) #batch 200 比 50的效果好 # batch(500) test accuracy 0.985714
if i % 500 == 0:
train_accuracy = accuracy.eval(feed_dict={
x: batch[0],
y_: batch[1],
keep_prob: 1.0
})
print('step %d, training accuracy %g' % (i, train_accuracy))
train_step.run(feed_dict={
x: batch[0],
y_: batch[1],
keep_prob: 0.5
})
print('test accuracy %g' % accuracy.eval(
feed_dict={
x: swallowsound.test.images,
y_: swallowsound.test.labels,
keep_prob: 1.0
}))
def main(_):
# Import data
# 注意one_hot参数表示我们要吧标签生成向量，比如目前的标签是数字“7”，分类数为10，则要生成一个[0,0,0,0,0,0,1,0,0,0]的向量
# 返回的结果是一个数据集，主要包括['train', 'validation', 'test']三个数据集，其中train和test的大小由原始数据确定，validation
# 从原始数据的train集合中选取，其大小要小于train的大小，具体大小由validation_size参数确定
# 数据集合中样本的分类数目由num_classes参数设置，标签label的值必须是[0，num_classes)之间的整数
num_classes = 2
# swallowsound = read_data_sets(FLAGS.data_dir,
# gzip_compress=False,
# train_imgaes='LearnSamples.bin',
# train_labels='LearnSamplesflag.bin',
# test_imgaes='TestSamples.bin',
# test_labels='TestSamplesflag.bin',
# one_hot=True,
# validation_size=50,
# num_classes = num_classes,
# MSB=False)
swallowsound = read_data_sets(FLAGS.data_dir,
gzip_compress=False,
train_imgaes='train-images-idx3-ubyte',
train_labels='train-labels-idx1-ubyte',
test_imgaes='t10k-images-idx3-ubyte',
test_labels='t10k-labels-idx1-ubyte',
one_hot=True,
validation_size=50,
num_classes=num_classes,
MSB=True)
original_shape = swallowsound.train.original_shape
if len(original_shape)<4:
return
size = original_shape[1]*original_shape[2]
# Create the model
x = tf.placeholder(tf.float32, [None, size])
W = tf.Variable(tf.zeros([size, num_classes]))
b = tf.Variable(tf.zeros([num_classes]))
y = tf.matmul(x, W) + b
# Define loss and optimizer
y_ = tf.placeholder(tf.float32, [None, num_classes])
# The raw formulation of cross-entropy,
#
# tf.reduce_mean(-tf.reduce_sum(y_ * tf.log(tf.nn.softmax(y)),
# reduction_indices=[1]))
#
# can be numerically unstable.
#
# So here we use tf.nn.softmax_cross_entropy_with_logits on the raw
# outputs of 'y', and then average across the batch.
cross_entropy = tf.reduce_mean(
tf.nn.softmax_cross_entropy_with_logits(labels=y_, logits=y))
train_step = tf.train.GradientDescentOptimizer(0.5).minimize(cross_entropy)
sess = tf.InteractiveSession()
tf.global_variables_initializer().run()
# Train
for _ in range(1000):
batch_xs, batch_ys = swallowsound.train.next_batch(100)
sess.run(train_step, feed_dict={x: batch_xs, y_: batch_ys})
# Test trained model
correct_prediction = tf.equal(tf.argmax(y, 1), tf.argmax(y_, 1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
print(sess.run(accuracy, feed_dict={x: swallowsound.test.images,
y_: swallowsound.test.labels}))
print("b: ",sess.run(b))