resnet50训练cifar10，请各位高手指正

使用resnet50从头训练cifar10，最终结果只有84%左右，貌似和 论文差很多，请各位高手指正。

首先加入cifar10的数据结构代码：

import cifar10,cifar10_input
import tensorflow as tf
import numpy as np
import time

#max_steps = 100000
max_steps = 100
data_dir = 'cifar-10-batches-bin'
batch_size = 128

# 配置每个 GPU 上占用的内存的比例
gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.95)
sess = tf.Session(config=tf.ConfigProto(gpu_options=gpu_options))

images_train, labels_train = cifar10_input.distorted_inputs(data_dir=data_dir,batch_size=batch_size)
images_test, labels_test = cifar10_input.inputs(eval_data=True,data_dir=data_dir,batch_size=batch_size)

sess = tf.InteractiveSession()
tf.global_variables_initializer().run()

tf.train.start_queue_runners()

start_time = time.time()
image_batch,label_batch = sess.run([images_train,labels_train])
duration = time.time() - start_time

print('Use Time = %.3f sec'%duration)

start_time = time.time()
image_batch,label_batch = sess.run([images_train,labels_train])
duration = time.time() - start_time

print('Use Time = %.3f sec'%duration)

加入resnet50的代码：

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

import tensorflow as tf

import resnet_utils


resnet_arg_scope = resnet_utils.resnet_arg_scope
slim = tf.contrib.slim


class NoOpScope(object):
  """No-op context manager."""

  def __enter__(self):
    return None

  def __exit__(self, exc_type, exc_value, traceback):
    return False


@slim.add_arg_scope
def bottleneck(inputs,
               depth,
               depth_bottleneck,
               stride,
               rate=1,
               outputs_collections=None,
               scope=None,
               use_bounded_activations=False):
  """Bottleneck residual unit variant with BN after convolutions.

  This is the original residual unit proposed in [1]. See Fig. 1(a) of [2] for
  its definition. Note that we use here the bottleneck variant which has an
  extra bottleneck layer.

  When putting together two consecutive ResNet blocks that use this unit, one
  should use stride = 2 in the last unit of the first block.

  Args:
    inputs: A tensor of size [batch, height, width, channels].
    depth: The depth of the ResNet unit output.
    depth_bottleneck: The depth of the bottleneck layers.
    stride: The ResNet unit's stride. Determines the amount of downsampling of
      the units output compared to its input.
    rate: An integer, rate for atrous convolution.
    outputs_collections: Collection to add the ResNet unit output.
    scope: Optional variable_scope.
    use_bounded_activations: Whether or not to use bounded activations. Bounded
      activations better lend themselves to quantized inference.

  Returns:
    The ResNet unit's output.
  """
  with tf.variable_scope(scope, 'bottleneck_v1', [inputs]) as sc:
    depth_in = slim.utils.last_dimension(inputs.get_shape(), min_rank=4)
    if depth == depth_in:
      shortcut = resnet_utils.subsample(inputs, stride, 'shortcut')
    else:
      shortcut = slim.conv2d(
          inputs,
          depth, [1, 1],
          stride=stride,
          activation_fn=tf.nn.relu6 if use_bounded_activations else None,
          scope='shortcut')

    residual = slim.conv2d(inputs, depth_bottleneck, [1, 1], stride=1,
                           scope='conv1')
    residual = resnet_utils.conv2d_same(residual, depth_bottleneck, 3, stride,
                                        rate=rate, scope='conv2')
    residual = slim.conv2d(residual, depth, [1, 1], stride=1,
                           activation_fn=None, scope='conv3')

    if use_bounded_activations:
      # Use clip_by_value to simulate bandpass activation.
      residual = tf.clip_by_value(residual, -6.0, 6.0)
      output = tf.nn.relu6(shortcut + residual)
    else:
      output = tf.nn.relu(shortcut + residual)

    return slim.utils.collect_named_outputs(outputs_collections,
                                            sc.name,
                                            output)

def resnet_v1(inputs,
              blocks,
              num_classes=None,
              is_training=True,
              global_pool=True,
              output_stride=None,
              include_root_block=True,
              spatial_squeeze=True,
              store_non_strided_activations=False,
              reuse=None,
              scope=None):
  """Generator for v1 ResNet models.

  This function generates a family of ResNet v1 models. See the resnet_v1_*()
  methods for specific model instantiations, obtained by selecting different
  block instantiations that produce ResNets of various depths.

  Training for image classification on Imagenet is usually done with [224, 224]
  inputs, resulting in [7, 7] feature maps at the output of the last ResNet
  block for the ResNets defined in [1] that have nominal stride equal to 32.
  However, for dense prediction tasks we advise that one uses inputs with
  spatial dimensions that are multiples of 32 plus 1, e.g., [321, 321]. In
  this case the feature maps at the ResNet output will have spatial shape
  [(height - 1) / output_stride + 1, (width - 1) / output_stride + 1]
  and corners exactly aligned with the input image corners, which greatly
  facilitates alignment of the features to the image. Using as input [225, 225]
  images results in [8, 8] feature maps at the output of the last ResNet block.

  For dense prediction tasks, the ResNet needs to run in fully-convolutional
  (FCN) mode and global_pool needs to be set to False. The ResNets in [1, 2] all
  have nominal stride equal to 32 and a good choice in FCN mode is to use
  output_stride=16 in order to increase the density of the computed features at
  small computational and memory overhead, cf. http://arxiv.org/abs/1606.00915.

  Args:
    inputs: A tensor of size [batch, height_in, width_in, channels].
    blocks: A list of length equal to the number of ResNet blocks. Each element
      is a resnet_utils.Block object describing the units in the block.
    num_classes: Number of predicted classes for classification tasks.
      If 0 or None, we return the features before the logit layer.
    is_training: whether batch_norm layers are in training mode. If this is set
      to None, the callers can specify slim.batch_norm's is_training parameter
      from an outer slim.arg_scope.
    global_pool: If True, we perform global average pooling before computing the
      logits. Set to True for image classification, False for dense prediction.
    output_stride: If None, then the output will be computed at the nominal
      network stride. If output_stride is not None, it specifies the requested
      ratio of input to output spatial resolution.
    include_root_block: If True, include the initial convolution followed by
      max-pooling, if False excludes it.
    spatial_squeeze: if True, logits is of shape [B, C], if false logits is
        of shape [B, 1, 1, C], where B is batch_size and C is number of classes.
        To use this parameter, the input images must be smaller than 300x300
        pixels, in which case the output logit layer does not contain spatial
        information and can be removed.
    store_non_strided_activations: If True, we compute non-strided (undecimated)
      activations at the last unit of each block and store them in the
      `outputs_collections` before subsampling them. This gives us access to
      higher resolution intermediate activations which are useful in some
      dense prediction problems but increases 4x the computation and memory cost
      at the last unit of each block.
    reuse: whether or not the network and its variables should be reused. To be
      able to reuse 'scope' must be given.
    scope: Optional variable_scope.

  Returns:
    net: A rank-4 tensor of size [batch, height_out, width_out, channels_out].
      If global_pool is False, then height_out and width_out are reduced by a
      factor of output_stride compared to the respective height_in and width_in,
      else both height_out and width_out equal one. If num_classes is 0 or None,
      then net is the output of the last ResNet block, potentially after global
      average pooling. If num_classes a non-zero integer, net contains the
      pre-softmax activations.
    end_points: A dictionary from components of the network to the corresponding
      activation.

  Raises:
    ValueError: If the target output_stride is not valid.
  """
  with tf.variable_scope(scope, 'resnet_v1', [inputs], reuse=reuse) as sc:
    end_points_collection = sc.original_name_scope + '_end_points'
    with slim.arg_scope([slim.conv2d, bottleneck,
                         resnet_utils.stack_blocks_dense],
                        outputs_collections=end_points_collection):
      with (slim.arg_scope([slim.batch_norm], is_training=is_training)
            if is_training is not None else NoOpScope()):
        net = inputs
        if include_root_block:
          if output_stride is not None:
            if output_stride % 4 != 0:
              raise ValueError('The output_stride needs to be a multiple of 4.')
            output_stride /= 4
          net = resnet_utils.conv2d_same(net, 64, 7, stride=2, scope='conv1')
          net = slim.max_pool2d(net, [3, 3], stride=2, scope='pool1')
        net = resnet_utils.stack_blocks_dense(net, blocks, output_stride,
                                              store_non_strided_activations)
        # Convert end_points_collection into a dictionary of end_points.
        end_points = slim.utils.convert_collection_to_dict(
            end_points_collection)

        if global_pool:
          # Global average pooling.
          net = tf.reduce_mean(net, [1, 2], name='pool5', keep_dims=True)
          end_points['global_pool'] = net
        if num_classes:
          net = slim.conv2d(net, num_classes, [1, 1], activation_fn=None,
                            normalizer_fn=None, scope='logits')
          end_points[sc.name + '/logits'] = net
          if spatial_squeeze:
            net = tf.squeeze(net, [1, 2], name='SpatialSqueeze')
            end_points[sc.name + '/spatial_squeeze'] = net
          end_points['predictions'] = slim.softmax(net, scope='predictions')
        return net, end_points
resnet_v1.default_image_size = 224

def resnet_v1_block(scope, base_depth, num_units, stride):
  """Helper function for creating a resnet_v1 bottleneck block.

  Args:
    scope: The scope of the block.
    base_depth: The depth of the bottleneck layer for each unit.
    num_units: The number of units in the block.
    stride: The stride of the block, implemented as a stride in the last unit.
      All other units have stride=1.

  Returns:
    A resnet_v1 bottleneck block.
  """
  return resnet_utils.Block(scope, bottleneck, [{
      'depth': base_depth * 4,
      'depth_bottleneck': base_depth,
      'stride': 1
  }] * (num_units - 1) + [{
      'depth': base_depth * 4,
      'depth_bottleneck': base_depth,
      'stride': stride
  }])


def resnet_v1_50(inputs,
                 num_classes=None,
                 is_training=True,
                 global_pool=True,
                 output_stride=None,
                 spatial_squeeze=True,
                 store_non_strided_activations=False,
                 reuse=None,
                 scope='resnet_v1_50'):
  """ResNet-50 model of [1]. See resnet_v1() for arg and return description."""
  blocks = [
      resnet_v1_block('block1', base_depth=64, num_units=3, stride=2),
      resnet_v1_block('block2', base_depth=128, num_units=4, stride=2),
      resnet_v1_block('block3', base_depth=256, num_units=6, stride=2),
      resnet_v1_block('block4', base_depth=512, num_units=3, stride=1),
  ]
  return resnet_v1(inputs, blocks, num_classes, is_training,
                   global_pool=global_pool, output_stride=output_stride,
                   include_root_block=True, spatial_squeeze=spatial_squeeze,
                   store_non_strided_activations=store_non_strided_activations,
                   reuse=reuse, scope=scope)

resnet_v1_50.default_image_size = 24

建立resnet50网络

height, width = 24, 24

image_holder = tf.placeholder(tf.float32, [batch_size, 24, 24, 3])
label_holder = tf.placeholder(tf.int32, [batch_size])

with slim.arg_scope(resnet_arg_scope()):
    net, end_points = resnet_v1_50(image_holder,10)
    print(end_points)

#看看网络情况
tf.global_variables_initializer().run()
see_net,see_end_points= sess.run([net,end_points],feed_dict={image_holder:image_batch,label_holder:label_batch})

print(see_net.shape)#应该是最后输出值
print(see_end_points)#应该是整个网络节点参数

定义loss函数，优化器：

def loss(logits, labels):
    labels = tf.cast(labels, tf.int64)
    cross_entropy = tf.nn.sparse_softmax_cross_entropy_with_logits(
        logits=logits, labels=labels, name='cross_entropy_per_example')
    cross_entropy_mean = tf.reduce_mean(cross_entropy, name='cross_entropy')
    tf.add_to_collection('losses', cross_entropy_mean)

    return cross_entropy_mean

gloss = loss(net, label_holder)

learnrate = 1e-3
train_op = tf.train.AdamOptimizer(learnrate).minimize(gloss)

#看看网络情况
tf.global_variables_initializer().run()
see_loss= sess.run([gloss],feed_dict={image_holder:image_batch,label_holder:label_batch})
print(see_loss)

训练网络：

import math
def SeePre():
    top_k_op = tf.nn.in_top_k(net, label_holder, 1)
    num_examples = 10000
    num_iter = int(math.ceil(num_examples / batch_size))
    true_count = 0  
    total_sample_count = num_iter * batch_size
    step = 0
    while step < num_iter:
        image_batch,label_batch = sess.run([images_test,labels_test])
        predictions = sess.run([top_k_op],feed_dict={image_holder: image_batch,
                                                 label_holder:label_batch})
        true_count += np.sum(predictions)
        step += 1
    precision = true_count / total_sample_count
    print('precision @ 1 = %.3f' % precision)

max_steps = 300000
SeePre()
for step in range(max_steps):
    start_time = time.time()
    image_batch,label_batch = sess.run([images_train,labels_train])
    _, nowloss = sess.run([train_op, gloss],feed_dict={image_holder: image_batch, 
                                                         label_holder:label_batch})
    duration = time.time() - start_time
    if step % 100 == 0:
        print('step = %d Use Time = %.3f sec loss=%.6f'%(step,duration,nowloss))
    if step % 1000 == 0:
        SeePre()
    if step % 10000 == 0:
        learnrate = 0.95*learnrate
SeePre()

最终结果0.84左右，和论文中说的94%相差很大啊，不知道以上代码是哪里出问题了呢？