Tf.slim Use Method __tensorflow

First to import Slim:

From Tensorflow.contrib Import Slim

Tf-slim mainly consists of the following:

Arg_scope
Data
Evluation
Layers
Learning
Losses
Metrics
Nets
Queues
Regularizers
Variables Layer

The most commonly used is the slim Layers, the creation of Layer is very convenient:

Input = ...
NET = slim.conv2d (input, 128, [3, 3], scope= ' conv1_1 ')
net = slim.max_pool2d (NET, kernel_size=[2,2], stride=2, scope= ' Pool1 ')
# generally (inputs=, kernel_size=, stride=, padding=?, ....)
NET = Slim.repeat (NET, 3, slim.conv2d, 256, [3, 3], scope= ' conv3 ')
# Repeat operation is to create a duplicate of a layer

Layer	Tf-slim
biasadd	slim.bias_add
batchnorm	slim.batch_norm
conv2d	slim.conv2d
conv2dinplane	slim.conv2d_i N_plane
conv2dtranspose (deconv)	slim.conv2d_transpose
F ullyconnected	slim.fully_connected
avgpool2d	slim.avg_pool2d
dropout	slim.dropout
flatten	slim.flatten
maxpool2d	slim.max_pool2d
onehotencoding	slim.one_hot_encoding
SeparableConv2	slim.separable_conv2d
unitnorm	slim.unit_norm

# All parameters are as follows
@add_arg_scope
def convolution2d_in_plane (
    inputs,
    kernel_size,
    stride=1,
    Padding= ' SAME ',
    Activation_fn=nn.relu,
    normalizer_fn=none,
    normalizer_params=none,
    weights_ Initializer=initializers.xavier_initializer (),
    Weights_regularizer=none,
    biases_initializer=init_ Ops.zeros_initializer (),
    Biases_regularizer=none,
    reuse=none,
    Variables_collections=none,
    Outputs_collections=none,
    trainable=true,
    scope=none):

Scopes:

The use of scope is also very convenient
After using Arg_scope, you can avoid adding duplicate parameters (such as weight_initializer=xxx, etc.) to each slim.conv2d.

# Add the same parameters to each slim.conv2d inside with
slim.arg_scope ([slim.conv2d], padding= ' SAME ',
                      weights_initializer= Tf.truncated_normal_initializer (stddev=0.01)
                      Weights_regularizer=slim.l2_regularizer (0.0005):
    net = SLIM.CONV2D (inputs, $, [one, one], scope= ' conv1 ')
    net = slim.conv2d (NET, 128, [one, one], padding= ' VALID ', scope= ' conv2 '
    net = slim.conv2d (NET, 256, [one, one], scope= ' conv3 ')

Variables:

Can be created as follows:

weights = slim.variable (' weights ',
                             shape=[10, 3, 3],
                             Initializer=tf.truncated_normal_initializer (StdDev =0.1),
                             Regularizer=slim.l2_regularizer (0.05),
                             device= '/cpu:0 ')

# Model Variables
weights = slim.model_variable (' weights ',
                              shape=[10, 3, 3],
                              initializer=tf.truncated _normal_initializer (stddev=0.1),
                              Regularizer=slim.l2_regularizer (0.05),
                              device= '/cpu:0 ')
model_ variables = Slim.get_model_variables ()

# Regular variables
my_var = slim.variable (' My_var ',
                       shape=[20, 1] ,
                       Initializer=tf.zeros_initializer ())
regular_variables_and_model_variables = Slim.get_variables ()

Training Loop

Simple training methods, call Slim.learning.create_train_op and Slim.learning.train to achieve

# call Slim.learning.create_train_op and Slim.learning.train to implement
g = tf. Graph ()

# Create The model and specify the losses
... ...

Total_loss = Slim.losses.get_total_loss ()
optimizer = Tf.train.GradientDescentOptimizer (learning_rate)

# CREATE_TRAIN_OP ensures that each time we ask for the loss, the Update_ops
# are run and the gradients being are applied too.
Train_op = Slim.learning.create_train_op (Total_loss, optimizer)
LogDir = ... # Where checkpoints are.

Slim.learning.train (
    train_op,
    logdir,
    number_of_steps=1000,
    save_summaries_secs=300,
    save_interval_secs=600):

other

There are other methods not written in this article, only the author (small white) used more listed
The other can refer to
Https://github.com/tensorflow/tensorflow/tree/master/tensorflow/contrib/slim Reference:

Tensorflow-slim:
Https://github.com/tensorflow/tensorflow/tree/master/tensorflow/contrib/slim